Skip to content
Back to search

PraisonAI

Private AI agents for highly regulated teams, and for anyone who wants AI on their own computer. Open-source Python agent framework that runs fully offline.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked yesterday

Check it deeper on your machine

It has changed since its scan on 19 Nov 2026.

As a Claude Code plugin, install it and then ask Claude Code:

/plugin marketplace add keithmackay/mackayi/plugin install scanskillsafety@mackayiDo a deep scan of https://github.com/MervinPraison/PraisonAI with scanskillsafety

Or from a terminal, clone it and run it:

git clone https://github.com/keithmackay/scanskillsafety-skill ~/.claude/skills/scanskillsafetynode ~/.claude/skills/scanskillsafety/cli.cjs --deep 'https://github.com/MervinPraison/PraisonAI'

It checks the exact commit: every file, its packages against OSV (api.osv.dev), how it compares with this page's scan (findsafeskills.com receives the repository's address), and, if you agree, Claude reading every file with your own Claude subscription. It never runs the code.

Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/mervinpraison--praisonai/badge.svg)](/listings/mervinpraison--praisonai)

HTML

<a href="/listings/mervinpraison--praisonai"><img src="/api/listings/mervinpraison--praisonai/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

Set a key for any supported provider (OpenAI shown here), or skip the key and use a local model as in the next section.

export OPENAI_API_KEY="your-api-key"
from praisonaiagents import Agent

agent = Agent(instructions="You are a helpful assistant.")
print(agent.start("In two sentences, why do some companies keep their data on their own computers?"))

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

README.md · README · 1,418 lines · 1 finding marked
  1. <p align="center">
  2. <picture>
  3. <source media="(prefers-color-scheme: dark)" srcset=".github/images/logo_dark.png" />
  4. <source media="(prefers-color-scheme: light)" srcset=".github/images/logo_light.png" />
  5. <img alt="PraisonAI Logo" src=".github/images/logo_light.png" width="250" />
  6. </picture>
  7. </p>
  8. <!-- mcp-name: io.github.MervinPraison/praisonai -->
  9. <p align="center">
  10. <a href="https://github.com/MervinPraison/PraisonAI"><img src="https://static.pepy.tech/badge/PraisonAI" alt="Total Downloads" /></a>
  11. <a href="https://github.com/MervinPraison/PraisonAI"><img src="https://img.shields.io/github/v/release/MervinPraison/PraisonAI" alt="Latest Stable Version" /></a>
  12. <a href="https://github.com/MervinPraison/PraisonAI"><img src="https://img.shields.io/badge/License-MIT-yellow.svg" alt="License" /></a>
  13. </p>
  14. # PraisonAI
  15. **Private AI that earns trust one task at a time.**
  16. PraisonAI runs AI agents on your own computers, even fully offline, so private data never has to leave the building.
  17. - **For highly regulated teams (early access):** each task moves to your own AI only after it passes a test. If quality slips, it moves back automatically, and every step is recorded for regulators. [How it works](#how-the-handover-works-early-access)
  18. - **For everyone else:** a free, open-source Python agent framework. With a local model you need no API key and pay no cloud bill, and your files stay yours.
  19. [2M+ downloads](https://pepy.tech/projects/praisonai) · [9.3K stars](https://github.com/MervinPraison/PraisonAI) · [600+ releases in the last year](https://github.com/MervinPraison/PraisonAI/releases) · [80+ contributors](https://github.com/MervinPraison/PraisonAI/graphs/contributors) · Python
  20. ```bash
  21. pip install praisonai
  22. ```
  23. Only need the Python library? `pip install praisonaiagents` installs just the lightweight core.
  24. ## Using PraisonAI for something else?
  25. PraisonAI is still a general-purpose, open-source agent framework, and everything works as before: the same packages, imports, commands and examples. Go straight to what you came for:
  26. - **Agents and multi-agent teams:** [Agents](https://praison.ai/docs/concepts/agents) · [Handoffs](https://praison.ai/docs/concepts/handoffs)
  27. - **Workflows:** [AgentFlow: route, parallel, loop, repeat](https://praison.ai/docs/concepts/agentflow)
  28. - **Memory and knowledge:** [Memory](https://praison.ai/docs/concepts/memory) · [Knowledge](https://praison.ai/docs/concepts/knowledge) · [RAG](https://praison.ai/docs/concepts/rag)
  29. - **Tools and MCP:** [Tools](https://praison.ai/docs/tools/tools) · [MCP](https://praison.ai/docs/mcp/transports)
  30. - **Bots (Telegram, Discord, Slack, WhatsApp):** [Bot CLI](https://praison.ai/docs/cli/bot) · [Bot gateway](https://praison.ai/docs/features/bot-gateway)
  31. - **Terminal CLI:** [`praisonai code`](https://praison.ai/docs/cli/code) · [`praisonai run`](https://praison.ai/docs/cli/run) · [`praisonai chat`](https://praison.ai/docs/cli/chat) · [CLI reference](https://praison.ai/docs/cli/cli-reference)
  32. - **Training and fine-tuning:** [PraisonAI Train](https://praison.ai/docs/train)
  33. - **Browser automation:** [Browser agent](https://praison.ai/docs/features/browser-agent) · [Browser CLI](https://praison.ai/docs/cli/browser)
  34. - **Sandboxes:** [Sandboxed code execution](https://praison.ai/docs/features/sandbox)
  35. - **Deploy:** [Deploy guide](https://praison.ai/docs/deploy)
  36. - **Dashboard and UI:** [Claw dashboard](https://praison.ai/docs/concepts/claw) · [Chat UI](https://praison.ai/docs/ui)
  37. - **YAML, no code:** [YAML agents](#-using-yaml-no-code) · [YAML reference](https://praison.ai/docs/features/yaml-configuration-reference)
  38. - **JavaScript and TypeScript:** [JavaScript SDK](https://praison.ai/docs/js)
  39. - **Desktop app:** [PraisonAI Desktop](https://praison.ai/docs/features/desktop/index) · [Downloads](#desktop-app)
  40. - **Everything else:** the [full feature reference](#full-feature-reference) further down this page
  41. ## Quick start
  42. Set a key for any supported provider (OpenAI shown here), or skip the key and use a local model as in the next section.
  43. ```bash
  44. export OPENAI_API_KEY="your-api-key"
  45. ```
  46. ```python
  47. from praisonaiagents import Agent
  48. agent = Agent(instructions="You are a helpful assistant.")
  49. print(agent.start("In two sentences, why do some companies keep their data on their own computers?"))
  50. ```
  51. ## Run fully offline
  52. Run the same agent on a model on your own computer with [Ollama](https://ollama.com). You need no API key, and your prompts and answers stay on the machine.
  53. Install the package and pull the model while you still have internet; after that, the agent runs with no network at all. If you installed only the core library, use `pip install "praisonaiagents[llm]"`, which adds the local-model support.
  54. ```bash
  55. # Start Ollama first (the app, or "ollama serve" in another terminal)
  56. ollama pull llama3.2
  57. export DO_NOT_TRACK=true # keeps telemetry off (PraisonAI sends none unless you turn it on)
  58. ```
  59. ```python
  60. from praisonaiagents import Agent
  61. agent = Agent(instructions="You are a helpful assistant.", model="ollama/llama3.2")
  62. print(agent.start("Why is the sky blue?"))
  63. ```
  64. If Ollama runs on another computer on your network, pass `base_url` (or set `OLLAMA_API_BASE`):
  65. ```python
  66. agent = Agent(instructions="...", model="ollama/llama3.2", base_url="http://192.168.1.10:11434")
  67. ```
  68. vLLM and LM Studio work the same way: use `hosted_vllm/` (vLLM) or `lm_studio/` with your server's `base_url`. More in the [Ollama guide](https://praison.ai/docs/models/ollama) and [Models docs](https://praison.ai/docs/models).
  69. ## How the handover works (early access)
  70. 1. **Watch:** records how each task is done today.
  71. 2. **Test:** your own AI takes an exam made from real past work it never studied.
  72. 3. **Hand over:** a task moves to your own AI only when it passes.
  73. 4. **Roll back:** if quality drops, the task moves back automatically.
  74. 5. **Record:** every handover leaves a record your regulator can check.
  75. **What is "your own AI"?** An open model, such as Llama, Qwen or Gemma, running on your own computers. Use it as it is, or train it on your own past work with [PraisonAI Train](https://praison.ai/docs/train) on your own GPU servers. Your team decides when to train; nothing trains automatically. Either way, a task moves to it only after it passes that exam.
  76. Highly regulated teams, such as banks, insurers and hospitals, want AI but can't send customer data to outside models. From 6 January 2027, OpenAI also takes no new self-serve fine-tuning jobs ([notice](https://developers.openai.com/api/docs/deprecations)), so teams that customise models need models they own. The handover lets these teams move work to AI they own, one task at a time, with a record that shows why each step was trusted.
  77. It is built on parts of this framework that are live today: tracing, evaluation, fine-tuning, model routing and local models. The handover itself is in early access, and we are taking on our first design partners. To take part, see [praison.ai/enterprise](https://praison.ai/enterprise/).
  78. ## Features
  79. - **Agents and multi-agent teams:** [Agents](https://praison.ai/docs/concepts/agents) · [Handoffs](https://praison.ai/docs/concepts/handoffs) · [Workflows](https://praison.ai/docs/concepts/agentflow)
  80. - **Memory:** [Memory](https://praison.ai/docs/concepts/memory) · [Knowledge and RAG](https://praison.ai/docs/concepts/rag) · [Sessions](https://praison.ai/docs/concepts/session-management)
  81. - **Tools and MCP:** [Tools](https://praison.ai/docs/tools/tools) · [MCP](https://praison.ai/docs/mcp/transports)
  82. - **Guardrails and approval:** [Guardrails](https://praison.ai/docs/concepts/guardrails) · [Human approval](https://praison.ai/docs/concepts/approval) · [Policy engine](https://praison.ai/docs/features/policy-engine)
  83. - **Tracing and evaluation:** [Tracing](https://praison.ai/docs/observability/overview) · [Telemetry](https://praison.ai/docs/features/telemetry) · [Langfuse](https://praison.ai/docs/observability/langfuse) · [Evaluation](https://praison.ai/docs/concepts/evaluation)
  84. - **Your choice of model:** [Models](https://praison.ai/docs/models) · [Model router](https://praison.ai/docs/features/model-router)
  85. - **Fine-tuning:** [PraisonAI Train](https://praison.ai/docs/train)
  86. - **Deploy:** [Deploy guide](https://praison.ai/docs/deploy)
  87. ## JavaScript and TypeScript
  88. ```bash
  89. npm install praisonai
  90. ```
  91. The TypeScript and JavaScript SDK has its own guide at [praison.ai/docs/js](https://praison.ai/docs/js). A short example is in [Using JavaScript Code](#-using-javascript-code) below.
  92. ## The company
  93. - **What we're building:** the task handover above, as a self-hosted product on top of this framework. It is in early access, and we are taking on our first design partners.
  94. - **Why now:** highly regulated teams want AI but can't send customer data to outside models, and from 6 January 2027 OpenAI takes no new self-serve fine-tuning jobs.
  95. - **How it fits together:** the framework stays free and MIT-licensed, and brings developers in. Regulated teams license the handover to run inside their own walls.
  96. - **Traction:** every number at the top of this page is public and linked to its source.
  97. - **Get in touch:** design partners, investors and press: enterprise@praison.ai or [praison.ai/enterprise](https://praison.ai/enterprise/)
  98. ## Docs, community and licence
  99. - **Documentation:** [praison.ai/docs](https://praison.ai/docs)
  100. - **Questions and bugs:** [Discussions](https://github.com/MervinPraison/PraisonAI/discussions) · [Issues](https://github.com/MervinPraison/PraisonAI/issues)
  101. - **Contributing:** [Contributing guide](https://github.com/MervinPraison/PraisonAI/blob/main/CONTRIBUTING.md)
  102. - **Videos and news:** [YouTube](https://youtube.com/@MervinPraison) · [X](https://x.com/MervinPraison) · [LinkedIn](https://linkedin.com/in/mervinpraison)
  103. - **Desktop app:** macOS, Windows and Linux builds are on the [latest release](https://github.com/MervinPraison/PraisonAI/releases/latest)
  104. - **Licence:** [MIT](https://github.com/MervinPraison/PraisonAI/blob/main/LICENSE)
  105. ---
  106. ## Full feature reference
  107. Everything below is the full reference that was on this page before. Nothing has been removed; the longest parts are folded, so click a heading's summary to open it.
  108. ### Desktop app
  109. <div align="center">
  110. <a href="https://github.com/MervinPraison/PraisonAI/releases/latest">
  111. <img src="https://img.shields.io/badge/Download_for_macOS-000000?style=for-the-badge&logo=apple&logoColor=white" alt="Download for macOS" />
  112. </a>
  113. &nbsp;&nbsp;
  114. <a href="https://github.com/MervinPraison/PraisonAI/releases/latest">
  115. <img src="https://img.shields.io/badge/Download_for_Windows-0078D6?style=for-the-badge&logo=windows&logoColor=white" alt="Download for Windows" />
  116. </a>
  117. &nbsp;&nbsp;
  118. <a href="https://github.com/MervinPraison/PraisonAI/releases/latest">
  119. <img src="https://img.shields.io/badge/Download_for_Linux-FCC624?style=for-the-badge&logo=linux&logoColor=black" alt="Download for Linux" />
  120. </a>
  121. </div>
  122. ### Other ways to install
  123. One-line installer:
  124. ```bash
  125. curl -fsSL https://praison.ai/install.sh | bash
    • Note: Pipes a downloaded script directly into a shell (curl | sh / wget | bash)
    • Line 173 on GitHub
  126. ```
  127. The full `praisonai` package, with the CLI, bots, dashboard and integrations:
  128. ```
  129. ██████╗ ██████╗ █████╗ ██╗███████╗ ██████╗ ███╗ ██╗ █████╗ ██╗
  130. ██╔══██╗██╔══██╗██╔══██╗██║██╔════╝██╔═══██╗████╗ ██║ ██╔══██╗██║
  131. ██████╔╝██████╔╝███████║██║███████╗██║ ██║██╔██╗ ██║ ███████║██║
  132. ██╔═══╝ ██╔══██╗██╔══██║██║╚════██║██║ ██║██║╚██╗██║ ██╔══██║██║
  133. ██║ ██║ ██║██║ ██║██║███████║╚██████╔╝██║ ╚████║ ██║ ██║██║
  134. ╚═╝ ╚═╝ ╚═╝╚═╝ ╚═╝╚═╝╚══════╝ ╚═════╝ ╚═╝ ╚═══╝ ╚═╝ ╚═╝╚═╝
  135. pip install praisonai
  136. ```
  137. <p align="center">
  138. <img src=".github/images/ui/claw-desktop-light.webp" alt="PraisonAI Claw dashboard in the Chat + Canvas layout: the chat on the left with a comparison table and a Python script the agent wrote, and the canvas on the right showing the table above the script" width="800" />
  139. </p>
  140. <div align="center">
  141. <a href="https://trendshift.io/repositories/9130" target="_blank"><img src="https://trendshift.io/api/badge/repositories/9130" alt="MervinPraison%2FPraisonAI | Trendshift" style="width: 250px; height: 55px;" width="250" height="55"/></a>
  142. </div>
  143. <div align="center">
  144. <a href="https://praison.ai/docs">
  145. <p align="center">
  146. <img src="https://img.shields.io/badge/📚_Documentation-Visit_praison.ai%2Fdocs-blue?style=for-the-badge&logo=bookstack&logoColor=white" alt="Documentation" />
  147. </p>
  148. </a>
  149. </div>
  150. ---
  151. ## 🎯 Use Cases
  152. AI agents solving real-world problems across industries:
  153. | Use Case | Description |
  154. |----------|-------------|
  155. | 🔍 **Research & Analysis** | Conduct deep research, gather information, and generate insights from multiple sources automatically |
  156. | 💻 **Code Generation** | Write, debug, and refactor code with AI agents that understand your codebase and requirements |
  157. | ✍️ **Content Creation** | Generate blog posts, documentation, marketing copy, and technical writing with multi-agent teams |
  158. | 📊 **Data Pipelines** | Extract, transform, and analyze data from APIs, databases, and web sources automatically |
  159. | 🤖 **Customer Support** | Deploy 24/7 support bots on Telegram, Discord, Slack with memory and knowledge-backed responses |
  160. | ⚙️ **Workflow Automation** | Automate multi-step business processes with agents that hand off tasks, verify results, and self-correct |
  161. ---
  162. ## 🚀 Meet your first Agent (Under 1 Minute)
  163. 1. Install the lightweight core SDK:
  164. ```bash
  165. pip install praisonaiagents
  166. export OPENAI_API_KEY="your-api-key"
  167. ```
  168. 2. Run your first autonomous agent:
  169. ```python
  170. from praisonaiagents import Agent
  171. # Give your agent a goal, and watch it work.
  172. agent = Agent(instructions="You are a senior data analyst.")
  173. print(agent.start("Analyze the top 3 tech trends of 2026 and format as a markdown table."))
  174. ```
  175. ---
  176. ## 🧬 The Five-Layer Agent Stack
  177. Most frameworks hand you one or two layers and leave the rest as homework. PraisonAI covers **all five** — plus the outer layer that decides *where* your agent actually runs.
  178. Each layer wraps the one inside it. When an agent misbehaves, the layer tells you where to look.
  179. <details>
  180. <summary><strong>Show the five layers, examples and Managed Agents</strong></summary>
  181. ```
  182. ┌─────────────────────────────────────────────────────────────────┐
  183. │ ⬡ MANAGED AGENTS — Where does it actually run? │
  184. │ ┌─────────────────────────────────────────────────────────────┐ │
  185. │ │ 5 · GRAPH — Who runs when, and who checks whom? │ │
  186. │ │ ┌─────────────────────────────────────────────────────────┐ │ │
  187. │ │ │ 4 · LOOP — When do we stop? │ │ │
  188. │ │ │ ┌─────────────────────────────────────────────────────┐ │ │ │
  189. │ │ │ │ 3 · HARNESS — Can it act, and be checked? │ │ │ │
  190. │ │ │ │ ┌─────────────────────────────────────────────────┐ │ │ │ │
  191. │ │ │ │ │ 2 · CONTEXT — Is the right thing in the window? │ │ │ │ │
  192. │ │ │ │ │ ┌─────────────────────────────────────────────┐ │ │ │ │ │
  193. │ │ │ │ │ │ 1 · PROMPT — Did I say it clearly? │ │ │ │ │ │
  194. │ │ │ │ │ └─────────────────────────────────────────────┘ │ │ │ │ │
  195. │ │ │ │ └─────────────────────────────────────────────────┘ │ │ │ │
  196. │ │ │ └─────────────────────────────────────────────────────┘ │ │ │
  197. │ │ └─────────────────────────────────────────────────────────┘ │ │
  198. │ └─────────────────────────────────────────────────────────────┘ │
  199. └─────────────────────────────────────────────────────────────────┘
  200. ```
  201. | Layer | The question it answers | PraisonAI |
  202. |:--|:--|:--|
  203. | **1 · Prompt** | Did I say it clearly? | `instructions=`, `role`/`goal`/`backstory`, `output=`, `templates=` |
  204. | **2 · Context** | Is the right thing in the window? | `memory=`, `knowledge=`, `context=`, handoff `ContextPolicy` |
  205. | **3 · Harness** | Can it act, and be checked? | `tools=`, `MCP()`, `guardrails=`, `approval=`, `hooks=`, `sandbox=` |
  206. | **4 · Loop** | When do we stop? | `execution=ExecutionConfig(...)`, `reflection=`, `autonomy=`, doom-loop detection |
  207. | **5 · Graph** | Who runs when, and who checks whom? | `AgentFlow`, `route()`, `parallel()`, `loop()`, `repeat()` |
  208. | **⬡ Managed** | *Where does it actually run?* | `tools_run_on="docker"` — one shared sandbox for the tools, or `run_on="anthropic"` for the whole agent |
  209. ### Layer 1 · Prompt — *Did I say it clearly?*
  210. Role, instructions, examples, output format.
  211. ```python
  212. from praisonaiagents import Agent
  213. agent = Agent(
  214. role="Senior Data Analyst",
  215. goal="Turn raw numbers into decisions",
  216. output="verbose", # markdown-formatted output
  217. )
  218. agent.start("Summarise Q3 revenue trends")
  219. ```
  220. ### Layer 2 · Context — *Is the right thing in the window?*
  221. Write, select, compress, isolate — the four context operations, one parameter each.
  222. ```python
  223. from praisonaiagents import Agent
  224. agent = Agent(
  225. instructions="You are a support engineer.",
  226. memory={"user_id": "u-42"}, # write — persists across runs (needs a user_id)
  227. knowledge=["docs/"], # select — retrieves only what's relevant
  228. context="summarize", # compress — auto-compacts before the limit
  229. )
  230. ```
  231. > **Isolate** is `handoffs=[specialist]` — a sub-agent inherits the last few messages and the intersection of your tools, not your whole transcript. [📖 Handoffs](https://praison.ai/docs/concepts/handoffs)
  232. ### Layer 3 · Harness — *Can it act, and be checked?*
  233. *Agent = Model + Harness.* Tool dispatch, plus the guides that steer before acting and the sensors that observe after.
  234. ```python
  235. from praisonaiagents import Agent, MCP, tool
  236. @tool
  237. def deploy(env: str) -> str:
  238. """Deploy the current build to an environment."""
  239. return f"Deployed to {env}"
  240. agent = Agent(
  241. name="ReleaseEngineer",
  242. instructions="You are a release engineer.",
  243. tools=[deploy, MCP("npx -y @modelcontextprotocol/server-filesystem /tmp")],
  244. approval=True, # guide — human gate before risky tools run
  245. )
  246. agent.start("Deploy to staging, then list the files you can read")
  247. ```
  248. ### Layer 4 · Loop — *When do we stop?*
  249. Hard iteration caps, budget ceilings, no-progress detection and completion checks — every brake is explicit.
  250. ```python
  251. from praisonaiagents import Agent, ExecutionConfig
  252. agent = Agent(
  253. instructions="Fix the failing tests.",
  254. execution=ExecutionConfig(max_iter=30, max_budget=0.50, on_budget_exceeded="stop"),
  255. autonomy=True, # required to drive the loop with run_autonomous()
  256. )
  257. result = agent.run_autonomous("Refactor the auth module", max_iterations=5)
  258. print(result.completion_reason)
  259. # goal | no_tool_calls | max_iterations | timeout | doom_loop | needs_help | error
  260. # (with on_budget_exceeded="stop", hitting the cap raises BudgetExceededError,
  261. # surfaced here as completion_reason="error")
  262. ```
  263. > **Doom-loop detection is on by default.** Repeated identical tool calls and A→B→A→B oscillation get caught — while a poller whose output keeps changing does not. [📖 Doom Loop Detection](https://praison.ai/docs/features/doom-loop-detection)
  264. ### Layer 5 · Graph — *Who runs when, and who checks whom?*
  265. Topology as a versionable artifact: prompt chaining, routing, parallelisation, orchestrator-worker.
  266. ```python
  267. from praisonaiagents import AgentFlow
  268. from praisonaiagents.workflows import route, parallel, repeat
  269. flow = AgentFlow(steps=[
  270. classifier,
  271. route({"bug": [bug_agent], "feature": [feature_agent], "default": [triage]}),
  272. parallel([reviewer, tester]), # fan out, join automatically
  273. repeat(editor, until=lambda ctx: "approved" in ctx.previous_result.lower(),
  274. max_iterations=3), # evaluator–optimizer
  275. ])
  276. flow.run("Ticket #123: login fails on Safari")
  277. ```
  278. > The same graph is expressible in YAML with no Python at all. [📖 AgentFlow](https://praison.ai/docs/concepts/agentflow)
  279. ### ⬡ Outside the stack: Managed Agents — *Where does it actually run?*
  280. The harness is commoditising; **where** the agent executes is the next multiplier. Rather than burning your laptop's CPU, hand an agent a short-lived cloud sandbox — repo, tools and tests run there.
  281. ```bash
  282. pip install praisonai
  283. ```
  284. The simplest way in is `tools_run_on=` — one whole team or workflow shares **one** sandbox, so a file written by step 1 is there for step 2. Thinking stays on your machine:
  285. ```python
  286. from praisonaiagents import Agent, AgentFlow
  287. writer = Agent(name="Writer", instructions="You write files.")
  288. reader = Agent(name="Reader", instructions="You read files.")
  289. flow = AgentFlow(tools_run_on="docker", steps=[writer, reader]) # or e2b | modal | daytona | flyio
  290. flow.run("Write 'hello' to /workspace/note.txt, then read it back")
  291. ```
  292. Same thing with no Python at all:
  293. ```yaml
  294. name: remote-demo
  295. tools_run_on: docker # every step shares one sandbox
  296. agents:
  297. writer: {role: Writer, goal: Write files}
  298. reader: {role: Reader, goal: Read files}
  299. steps:
  300. - agent: writer
  301. action: "Write 'hello' to /workspace/note.txt"
  302. - agent: reader
  303. action: "Read /workspace/note.txt"
  304. ```
  305. For a single agent, two words cover it — and they answer different questions:
  306. ```python
  307. from praisonaiagents import Agent
  308. # A. Only the TOOLS move. Thinking stays on your machine.
  309. agent = Agent(name="builder", instructions="You build things.",
  310. tools_run_on="docker") # docker | e2b | modal | daytona | flyio
  311. # tenki | sandlock | ssh | novita | subprocess
  312. # B. The WHOLE agent moves — model calls, loop and tools
  313. agent = Agent(name="teacher", instructions="You teach.", run_on="anthropic") # hosted
  314. agent = Agent(name="builder", instructions="You build.", run_on="docker") # self-hosted
  315. agent.start("Write a Python script that prints the first 10 primes, then run it")
  316. ```
  317. Ask any object where it runs, and it will tell you:
  318. ```python
  319. >>> Agent(name="builder", instructions="x", tools_run_on="docker")
  320. Agent(name='builder', thinks_on='this machine', tools_run_on='a Docker container')
  321. >>> agent.where_does_it_run()
  322. Thinking (the AI model calls) happens on this machine.
  323. Tools run on a Docker container.
  324. Your own tools (check_db) still run on this machine -- only shell, file and
  325. code tools move. They read and write this machine's files.
  326. ```
  327. Naming a place that cannot do the job is a typo, not a preference, so it says so:
  328. ```python
  329. >>> Agent(name="x", instructions="i", run_on="e2b")
  330. TypeError: Agent(run_on='e2b') is not valid: run_on= places the whole agent
  331. -- model calls, loop and tools -- on a managed runtime, and 'e2b' runs
  332. commands but cannot host an agent loop.
  333. To run only the tools there: Agent(tools_run_on='e2b')
  334. ```
  335. To run one block of code somewhere else, name the place on that call:
  336. ```python
  337. agent.execute_code_sync("print(6 * 7)", run_in="sandlock") # kernel-enforced
  338. ```
  339. See what is running and reclaim strays:
  340. ```bash
  341. praisonai managed ps # list running sandboxes
  342. praisonai managed stop --all # reclaim them
  343. ```
  344. Sandboxes shut themselves down when idle (`auto_shutdown`, `idle_timeout_s`), and a post-setup snapshot is reused so the next run skips the image pull and dependency install. Commit a `.praisonai/environment.yaml` and the environment travels with the repo.
  345. > 📖 [20 runnable examples](examples/python/managed-agents/) · manage sessions with `praisonai managed sessions list <agent-id>` or `praisonai managed sessions resume <session-id> "<prompt>"`
  346. <sub>Stack framing adapted from [The Five-Layer Agent Stack](https://mer.vin/2026/07/five-layer-agent-stack-match-bug-to-right-layer/) and [Agent Harnesses vs Orbs](https://mer.vin/2026/08/agent-harnesses-vs-orbs-why-remote-sandboxes-beat-local-agent-loops/).</sub>
  347. </details>
  348. ---
  349. ## 🌌 The PraisonAI Ecosystem
  350. Start simple with the core SDK, or expand to full visual builders and dashboards when you're ready.
  351. * **Core SDK (`praisonaiagents`)**: For pure Python development. `pip install praisonaiagents`
  352. * 💻 **PraisonAI CLI (`praisonai`)**: For terminal-based developers. `pip install praisonai`
  353. * 🦞 **Claw Dashboard**: Connect agents directly to Telegram, Slack, or Discord. `pip install "praisonai[claw]"`
  354. * 🔗 **Flow Visual Builder**: Drag-and-drop workflow creation. `pip install "praisonai[flow]"`
  355. * 🤖 **PraisonAI UI**: Chat with a canvas for the tables and code your agents write. `pip install "praisonai[ui]"`
  356. ### JavaScript SDK
  357. ```bash
  358. npm install praisonai
  359. ```
  360. ## 🧠 Supported Providers & Features
  361. Powered by 100+ LLMs (OpenAI, Anthropic, Gemini & local models).
  362. <p align="center">
  363. <img src="https://img.shields.io/badge/OpenAI-412991?style=flat&logo=openai&logoColor=white" alt="OpenAI" />
  364. <img src="https://img.shields.io/badge/Anthropic-191919?style=flat&logo=anthropic&logoColor=white" alt="Anthropic" />
  365. <img src="https://img.shields.io/badge/Google_Gemini-4285F4?style=flat&logo=google&logoColor=white" alt="Google Gemini" />
  366. <img src="https://img.shields.io/badge/DeepSeek-566AB2?style=flat" alt="DeepSeek" />
  367. <img src="https://img.shields.io/badge/Azure-0078D4?style=flat&logo=microsoftazure&logoColor=white" alt="Azure" />
  368. <img src="https://img.shields.io/badge/Ollama-000000?style=flat" alt="Ollama" />
  369. <img src="https://img.shields.io/badge/Groq-F05237?style=flat" alt="Groq" />
  370. <img src="https://img.shields.io/badge/Mistral-FF7000?style=flat" alt="Mistral" />
  371. <img src="https://img.shields.io/badge/Cerebras-F05A28?style=flat" alt="Cerebras" />
  372. <img src="https://img.shields.io/badge/Cohere-39594D?style=flat" alt="Cohere" />
  373. <img src="https://img.shields.io/badge/OpenRouter-6467F2?style=flat" alt="OpenRouter" />
  374. <img src="https://img.shields.io/badge/Perplexity-20808D?style=flat" alt="Perplexity" />
  375. <img src="https://img.shields.io/badge/Fireworks-FF6B35?style=flat" alt="Fireworks" />
  376. <img src="https://img.shields.io/badge/AWS_Bedrock-FF9900?style=flat&logo=amazonaws&logoColor=white" alt="AWS Bedrock" />
  377. <img src="https://img.shields.io/badge/xAI_Grok-000000?style=flat" alt="xAI Grok" />
  378. <img src="https://img.shields.io/badge/Vertex_AI-4285F4?style=flat&logo=googlecloud&logoColor=white" alt="Vertex AI" />
  379. <img src="https://img.shields.io/badge/HuggingFace-FFD21E?style=flat&logo=huggingface&logoColor=black" alt="HuggingFace" />
  380. <img src="https://img.shields.io/badge/Together_AI-000000?style=flat" alt="Together AI" />
  381. <img src="https://img.shields.io/badge/Databricks-FF3621?style=flat&logo=databricks&logoColor=white" alt="Databricks" />
  382. <img src="https://img.shields.io/badge/Replicate-262626?style=flat" alt="Replicate" />
  383. <img src="https://img.shields.io/badge/Cloudflare-F38020?style=flat&logo=cloudflare&logoColor=white" alt="Cloudflare" />
  384. </p>
  385. <details>
  386. <summary><strong>View all 24 providers with examples</strong></summary>
  387. | Provider | Example |
  388. |----------|:-------:|
  389. | OpenAI | [Example](examples/python/providers/openai/openai_gpt4_example.py) |
  390. | Anthropic | [Example](examples/python/providers/anthropic/anthropic_claude_example.py) |
  391. | Google Gemini | [Example](examples/python/providers/google/google_gemini_example.py) |
  392. | Ollama | [Example](examples/python/providers/ollama/ollama-agents.py) |
  393. | Groq | [Example](examples/python/providers/groq/kimi_with_groq_example.py) |
  394. | DeepSeek | [Example](examples/python/providers/deepseek/deepseek_example.py) |
  395. | xAI Grok | [Example](examples/python/providers/xai/xai_grok_example.py) |
  396. | Mistral | [Example](examples/python/providers/mistral/mistral_example.py) |
  397. | Cohere | [Example](examples/python/providers/cohere/cohere_example.py) |
  398. | Perplexity | [Example](examples/python/providers/perplexity/perplexity_example.py) |
  399. | Fireworks | [Example](examples/python/providers/fireworks/fireworks_example.py) |
  400. | Together AI | [Example](examples/python/providers/together/together_ai_example.py) |
  401. | OpenRouter | [Example](examples/python/providers/openrouter/openrouter_example.py) |
  402. | HuggingFace | [Example](examples/python/providers/huggingface/huggingface_example.py) |
  403. | Azure OpenAI | [Example](examples/python/providers/azure/azure_openai_example.py) |
  404. | AWS Bedrock | [Example](examples/python/providers/aws/aws_bedrock_example.py) |
  405. | Google Vertex | [Example](examples/python/providers/vertex/vertex_example.py) |
  406. | Databricks | [Example](examples/python/providers/databricks/databricks_example.py) |
  407. | Cloudflare | [Example](examples/python/providers/cloudflare/cloudflare_example.py) |
  408. | AI21 | [Example](examples/python/providers/ai21/ai21_example.py) |
  409. | Replicate | [Example](examples/python/providers/replicate/replicate_example.py) |
  410. | SageMaker | [Example](examples/python/providers/sagemaker/sagemaker_example.py) |
  411. | Moonshot | [Example](examples/python/providers/moonshot/moonshot_example.py) |
  412. | vLLM | [Example](examples/python/providers/vllm/vllm_example.py) |
  413. </details>
  414. <div align="center">
  415. <a href="https://x.com/elonmusk/status/1893870468249141688" target="_blank">
  416. <img src="https://img.shields.io/badge/Highlighted_by_Elon_Musk-000000?style=for-the-badge&logo=x&logoColor=white" alt="Highlighted by Elon Musk" />
  417. </a>
  418. <br>
  419. <br>
  420. <a href="https://x.com/elonmusk/status/1893870468249141688" target="_blank">
  421. <img src=".github/images/elon_musk_praisonai.png" alt="Highlighted by Elon Musk" width="600" />
  422. </a>
  423. <p><em>"Grok 3 customer support" — <a href="https://x.com/elonmusk/status/1893870468249141688">Elon Musk quoting PraisonAI's tutorial</a></em></p>
  424. </div>
  425. <br>
  426. ---
  427. ## 🌟 Why PraisonAI?
  428. | | Feature | How |
  429. |--|---------|-----|
  430. | 🔌 | **MCP Protocol** — stdio, HTTP, WebSocket, SSE | `tools=MCP("npx ...")` |
  431. | 🧠 | **Planning Mode** — plan → execute → reason | `planning=True` |
  432. | 🔍 | **Deep Research** — multi-step autonomous research | [Docs](https://praison.ai/docs/agents/deep-research) |
  433. | 🤖 | **External Agents** — orchestrate Claude Code, Gemini CLI, Codex | [Docs](https://praison.ai/docs/code/external-agents) |
  434. | 🔄 | **Agent Handoffs** — seamless conversation passing | `handoffs=[other_agent]` |
  435. | 🛡️ | **Guardrails** — input/output validation | [Docs](https://praison.ai/docs/concepts/guardrails) |
  436. | | **Web Search + Fetch** — native browsing | `web=True` |
  437. | 🪞 | **Self Reflection** — agent reviews its own output | [Docs](https://praison.ai/docs/concepts/reflection) |
  438. | 🔀 | **Workflow Patterns** — route, parallel, loop, repeat | [Docs](https://praison.ai/docs/concepts/agentflow) |
  439. | 🧠 | **Memory (zero deps)** — works out of the box | `memory=True` |
  440. <details>
  441. <summary><strong>View all 25 features</strong></summary>
  442. | | Feature | How |
  443. |--|---------|-----|
  444. | 💡 | **Prompt Caching** — reduce latency + cost | `caching=True` |
  445. | 💾 | **Sessions + Auto-Save** — persistent state across restarts | `auto_save="my-project"` |
  446. | 💭 | **Thinking Budgets** — control reasoning depth | `agent.thinking_budget = 1024` |
  447. | 📚 | **RAG + Quality-Based RAG** — auto quality scoring retrieval | [Docs](https://praison.ai/docs/concepts/rag) |
  448. | 📊 | **Model Router** — auto-routes to cheapest capable model | [Docs](https://praison.ai/docs/features/model-router) |
  449. | 🧊 | **Shadow Git Checkpoints** — auto-rollback on failure | [Docs](https://praison.ai/docs/features/checkpoints) |
  450. | 📡 | **A2A Protocol** — agent-to-agent interop | [Docs](https://praison.ai/docs/features/a2a) |
  451. | 📏 | **Context Compaction** — never hit token limits | [Docs](https://praison.ai/docs/features/context-compaction) |
  452. | 📡 | **Telemetry** — OpenTelemetry traces, spans, metrics | [Docs](https://praison.ai/docs/features/telemetry) |
  453. | 📜 | **Policy Engine** — declarative agent behavior control | [Docs](https://praison.ai/docs/features/policy-engine) |
  454. | 🔄 | **Background Tasks** — fire-and-forget agents | [Docs](https://praison.ai/docs/features/background-tasks) |
  455. | 🔁 | **Doom Loop Detection** — auto-recovery from stuck agents | [Docs](https://praison.ai/docs/features/doom-loop-detection) |
  456. | 🕸️ | **Graph Memory** — Neo4j-style relationship tracking | [Docs](https://praison.ai/docs/features/graph-memory) |
  457. | 🏖️ | **Sandbox Execution** — isolated code execution | [Docs](https://praison.ai/docs/features/sandbox) |
  458. | 🖥️ | **Bot Gateway** — multi-agent routing across channels | [Docs](https://praison.ai/docs/features/bot-gateway) |
  459. </details>
  460. ---
  461. ## 📘 Using Python Code
  462. <details>
  463. <summary><strong>Show Python examples: single agent, multi-agent, MCP, tools, databases, Claw, Langflow, UI</strong></summary>
  464. > Unsure whether a result is a bug or model randomness? Check the [SDK Behavior Matrix](https://github.com/MervinPraison/PraisonAI/blob/main/src/praisonai-agents/docs/sdk-behavior-matrix.md) for `stream`, memory prefetch, structured output, teams and AG-UI.
  465. ### 1. Single Agent
  466. ```python
  467. from praisonaiagents import Agent
  468. agent = Agent(instructions="You are a helpful AI assistant")
  469. agent.start("Write a movie script about a robot in Mars")
  470. ```
  471. ### 2. Multi Agents
  472. ```python
  473. from praisonaiagents import Agent, Agents
  474. research_agent = Agent(instructions="Research about AI")
  475. summarise_agent = Agent(instructions="Summarise research agent's findings")
  476. agents = Agents(agents=[research_agent, summarise_agent])
  477. agents.start()
  478. ```
  479. ### 3. MCP (Model Context Protocol)
  480. ```python
  481. from praisonaiagents import Agent, MCP
  482. # stdio - Local NPX/Python servers
  483. agent = Agent(tools=MCP("npx @modelcontextprotocol/server-memory"))
  484. # Streamable HTTP - Production servers
  485. agent = Agent(tools=MCP("https://api.example.com/mcp"))
  486. # WebSocket - Real-time bidirectional
  487. agent = Agent(tools=MCP("wss://api.example.com/mcp", auth_token="token"))
  488. # With environment variables
  489. agent = Agent(
  490. tools=MCP(
  491. command="npx",
  492. args=["-y", "@modelcontextprotocol/server-brave-search"],
  493. env={"BRAVE_API_KEY": "your-key"}
  494. )
  495. )
  496. ```
  497. > 📖 [Full MCP docs](https://praison.ai/docs/mcp/transports) — stdio, HTTP, WebSocket, SSE transports
  498. ### 4. Custom Tools
  499. ```python
  500. from praisonaiagents import Agent, tool
  501. @tool
  502. def search(query: str) -> str:
  503. """Search the web for information."""
  504. return f"Results for: {query}"
  505. @tool
  506. def calculate(expression: str) -> float:
  507. """Safely evaluate a numeric arithmetic expression."""
  508. import ast
  509. import operator
  510. # Define allowed operations
  511. _OPS = {
  512. ast.Add: operator.add,
  513. ast.Sub: operator.sub,
  514. ast.Mult: operator.mul,
  515. ast.Div: operator.truediv,
  516. ast.Pow: operator.pow,
  517. ast.USub: operator.neg,
  518. ast.UAdd: operator.pos,
  519. }
  520. def _safe_eval(node):
  521. if isinstance(node, ast.Constant) and isinstance(node.value, (int, float)):
  522. return node.value
  523. elif isinstance(node, ast.BinOp) and type(node.op) in _OPS:
  524. return _OPS[type(node.op)](_safe_eval(node.left), _safe_eval(node.right))
  525. elif isinstance(node, ast.UnaryOp) and type(node.op) in _OPS:
  526. return _OPS[type(node.op)](_safe_eval(node.operand))
  527. else:
  528. raise ValueError("Unsupported expression")
  529. try:
  530. return _safe_eval(ast.parse(expression, mode="eval").body)
  531. except (ValueError, SyntaxError, TypeError, ZeroDivisionError, OverflowError):
  532. raise ValueError("Invalid arithmetic expression")
  533. agent = Agent(
  534. instructions="You are a helpful assistant",
  535. tools=[search, calculate]
  536. )
  537. agent.start("Search for AI news and calculate 15*4")
  538. ```
  539. > ⚠️ **Security Note:** Never use `eval()`, `exec()`, or `subprocess` in tool functions that process LLM-generated or user-supplied input. Always validate and sanitize inputs to prevent code injection attacks.
  540. > 📖 [Full tools docs](https://praison.ai/docs/tools/tools) — BaseTool, tool packages, 100+ built-in tools
  541. ### 5. Persistence (Databases)
  542. ```python
  543. from praisonaiagents import Agent, db
  544. agent = Agent(
  545. name="Assistant",
  546. memory={
  547. "db": db(database_url="postgresql://localhost/mydb"),
  548. "session_id": "my-session",
  549. },
  550. )
  551. agent.chat("Hello!") # Auto-persists messages, runs, traces
  552. ```
  553. > 📖 [Full persistence docs](https://praison.ai/docs/databases/overview) — PostgreSQL, MySQL, SQLite, MongoDB, Redis, and 20+ more
  554. ### 6. PraisonAI Claw 🦞 (Dashboard UI)
  555. Connect your AI agents to **Telegram, Discord, Slack, WhatsApp** and more — all from a single command.
  556. ```bash
  557. pip install "praisonai[claw]"
  558. praisonai claw
  559. ```
  560. #### Required Environment Variables
  561. Copy `.env.example` to `.env` and configure the following variables:
  562. | Variable | Required | Description |
  563. |----------|----------|-------------|
  564. | `OPENAI_API_KEY` | Yes | OpenAI API key for all LLM calls |
  565. | `TAVILY_API_KEY` | Yes (Claw) | Tavily key for the built-in web-search tool. Get one free at https://app.tavily.com |
  566. Open **http://localhost:8082** — the dashboard opens in **Chat + Canvas**: the chat on the left, and the tables, code and files your agents write on a canvas beside it (sort, copy, download, compare versions). Every other page (Agents, Memory, Knowledge, Channels, Guardrails, Cron, and more) is in the sidebar or the command palette (⌘K). Add messaging channels directly from the UI.
  567. > 📖 [Full Claw docs](https://praison.ai/docs/concepts/claw) — platform tokens, CLI options, Docker, and YAML agent mode
  568. ### 7. Langflow Integration 🔗 (Visual Flow Builder)
  569. Build multi-agent workflows visually with **drag-and-drop** components in Langflow.
  570. ```bash
  571. pip install "praisonai[flow]"
  572. praisonai flow
  573. ```
  574. Open **http://localhost:7861** — use the **Agent** and **Agent Team** components to create sequential or parallel workflows. Connect Chat Input → Agent Team → Chat Output for instant multi-agent pipelines.
  575. > 📖 [Full Flow docs](https://praison.ai/docs/concepts/agentflow) — visual agent building, component reference, and deployment
  576. ### 8. PraisonAI UI 🤖 (Chat + Canvas)
  577. Chat with your AI agents; the tables and code they write open on a canvas beside the chat.
  578. ```bash
  579. pip install "praisonai[ui]"
  580. praisonai ui # chat → http://localhost:8081
  581. praisonai ui agents # YAML agents
  582. praisonai ui bot # bot interface
  583. praisonai ui realtime # realtime voice
  584. ```
  585. <p align="center">
  586. <img src=".github/images/ui/ui-welcome-desktop-light.webp" alt="praisonai ui opening a new chat: the PraisonAI Chat title, the app's welcome greeting under it, four starter prompts and the message box" width="400" />
  587. <img src=".github/images/ui/ui-desktop-dark.webp" alt="praisonai ui in dark mode: the chat on the left and the canvas with the comparison table and the benchmark script on the right" width="400" />
  588. </p>
  589. - **Welcome:** each app greets you in the new-chat hero, in every layout. Set the text with `UIPreset(welcome=...)` or your own `@aiui.welcome` handler.
  590. - **Layout:** Chat + Canvas by default. Pick another for one run with `AIUI_LAYOUT=chat` (no canvas), `workbench`, `command`, `guided` or `mission`, or switch from the command palette (⌘K).
  591. - **Look:** light and dark follow the toggle in the top bar; ⌘K → **Switch look: Pro** swaps the Paper look for the plain one.
  592. - **Stop:** the Stop button cancels the agent's reply. A reply that sends nothing for 10 minutes is cancelled too (`AIUI_REPLY_IDLE_TIMEOUT`, in seconds; `0` turns it off).
  593. - **Uploads:** request bodies are capped at 25 MiB (`AIUI_MAX_BODY_MB`; `0` turns it off).
  594. - **Data:** chat history is saved by PraisonAI under `~/.praisonai/sessions` (set `PRAISONAI_HOME` to move it); the UI's own settings live in `~/.praisonaiui` (set `AIUI_DATA_DIR` to move them). To keep both on a mounted volume, set both variables.
  595. - An existing `~/.praisonai/claw/app.py` from an older install keeps aiui's own default layout (Chat, no canvas) and has no welcome greeting; delete it to get the new default app, or add `layout="chat-canvas"` to its `configure_host(...)` call.
  596. </details>
  597. ---
  598. ## 📄 Using YAML (No Code)
  599. <details>
  600. <summary><strong>Show YAML examples</strong></summary>
  601. ### Example 1: Two Agents Working Together
  602. Create `agents.yaml`:
  603. ```yaml
  604. framework: praisonai
  605. topic: "Write a blog post about AI"
  606. agents:
  607. researcher:
  608. role: Research Analyst
  609. goal: Research AI trends and gather information
  610. instructions: "Find accurate information about AI trends"
  611. writer:
  612. role: Content Writer
  613. goal: Write engaging blog posts
  614. instructions: "Write clear, engaging content based on research"
  615. ```
  616. Run with:
  617. ```bash
  618. praisonai agents.yaml
  619. ```
  620. > The agents automatically work together sequentially
  621. ### Example 2: Agent with Custom Tool
  622. Create two files in the same folder:
  623. **agents.yaml:**
  624. ```yaml
  625. framework: praisonai
  626. topic: "Calculate the sum of 25 and 15"
  627. agents:
  628. calculator_agent:
  629. role: Calculator
  630. goal: Perform calculations
  631. instructions: "Use the add_numbers tool to help with calculations"
  632. tools:
  633. - add_numbers
  634. ```
  635. **tools.py:**
  636. ```python
  637. def add_numbers(a: float, b: float) -> float:
  638. """
  639. Add two numbers together.
  640. Args:
  641. a: First number
  642. b: Second number
  643. Returns:
  644. The sum of a and b
  645. """
  646. return a + b
  647. ```
  648. Run with:
  649. ```bash
  650. praisonai agents.yaml
  651. ```
  652. > 💡 **Tips:**
  653. > - Use the function name (e.g., `add_numbers`) in the tools list, not the file name
  654. > - Tools in `tools.py` are automatically discovered
  655. > - The function's docstring helps the AI understand how to use it
  656. </details>
  657. ---
  658. ## 🎯 CLI Quick Reference
  659. | Category | Commands |
  660. |----------|----------|
  661. | **Execution** | `praisonai`, `--auto`, `--interactive`, `--chat` |
  662. | **Research** | `research`, `--query-rewrite`, `--deep-research` |
  663. | **Planning** | `--planning`, `--planning-tools`, `--planning-reasoning` |
  664. | **Workflows** | `workflow run`, `workflow list`, `workflow auto` |
  665. | **Memory** | `memory show`, `memory add`, `memory search`, `memory clear` |
  666. | **Knowledge** | `knowledge add`, `knowledge query`, `knowledge list` |
  667. | **Sessions** | `session list`, `session resume`, `session delete` |
  668. | **Tools** | `tools list`, `tools info`, `tools search` |
  669. | **MCP** | `mcp list`, `mcp create`, `mcp enable` |
  670. | **Development** | `commit`, `docs`, `checkpoint`, `hooks` |
  671. | **Scheduling** | `schedule start`, `schedule list`, `schedule stop` |
  672. > 📖 [Full CLI reference](https://praison.ai/docs/cli/cli-reference)
  673. ---
  674. ## ✨ Key Features
  675. <details>
  676. <summary><strong>Show all features by category</strong></summary>
  677. <details open>
  678. <summary><strong>🤖 Core Agents</strong></summary>
  679. | Feature | Code | Docs |
  680. |---------|:----:|:----:|
  681. | Single Agent | [Example](examples/python/agents/single-agent.py) | [📖](https://praison.ai/docs/agents/single) |
  682. | Multi Agents | [Example](examples/python/general/mini_agents_example.py) | [📖](https://praison.ai/docs/concepts/agents) |
  683. | Auto Agents | [Example](examples/python/general/auto_agents_example.py) | [📖](https://praison.ai/docs/features/autoagents) |
  684. | Self Reflection AI Agents | [Example](examples/python/concepts/self-reflection-details.py) | [📖](https://praison.ai/docs/concepts/reflection) |
  685. | Reasoning AI Agents | [Example](examples/python/concepts/reasoning-extraction.py) | [📖](https://praison.ai/docs/features/reasoning) |
  686. | Multi Modal AI Agents | [Example](examples/python/general/multimodal.py) | [📖](https://praison.ai/docs/features/multimodal) |
  687. </details>
  688. <details>
  689. <summary><strong>🔄 Workflows</strong></summary>
  690. | Feature | Code | Docs |
  691. |---------|:----:|:----:|
  692. | Simple Workflow | [Example](examples/python/workflows/simple_workflow.py) | [📖](https://praison.ai/docs/concepts/agentflow) |
  693. | Workflow with Agents | [Example](examples/python/workflows/workflow_with_agents.py) | [📖](https://praison.ai/docs/concepts/agentflow) |
  694. | Agentic Routing (`route()`) | [Example](examples/python/workflows/workflow_routing.py) | [📖](https://praison.ai/docs/features/routing) |
  695. | Parallel Execution (`parallel()`) | [Example](examples/python/workflows/workflow_parallel.py) | [📖](https://praison.ai/docs/features/parallelisation) |
  696. | Loop over List/CSV (`loop()`) | [Example](examples/python/workflows/workflow_loop_csv.py) | [📖](https://praison.ai/docs/features/repetitive) |
  697. | Evaluator-Optimizer (`repeat()`) | [Example](examples/python/workflows/workflow_repeat.py) | [📖](https://praison.ai/docs/concepts/evaluation) |
  698. | Conditional Steps | [Example](examples/python/workflows/workflow_conditional.py) | [📖](https://praison.ai/docs/concepts/agentflow) |
  699. | Workflow Branching | [Example](examples/python/workflows/workflow_branching.py) | [📖](https://praison.ai/docs/concepts/agentflow) |
  700. | Workflow Early Stop | [Example](examples/python/workflows/workflow_early_stop.py) | [📖](https://praison.ai/docs/concepts/agentflow) |
  701. | Workflow Checkpoints | [Example](examples/python/workflows/workflow_checkpoints.py) | [📖](https://praison.ai/docs/concepts/agentflow) |
  702. </details>
  703. <details>
  704. <summary><strong>💻 Code & Development</strong></summary>
  705. | Feature | Code | Docs |
  706. |---------|:----:|:----:|
  707. | Code Interpreter Agents | [Example](examples/python/agents/code-agent.py) | [📖](https://praison.ai/docs/features/codeagent) |
  708. | AI Code Editing Tools | [Example](examples/python/code/code_editing_example.py) | [📖](https://praison.ai/docs/code/editing) |
  709. | External Agents (All) | [Example](examples/python/code/external_agents_example.py) | [📖](https://praison.ai/docs/code/external-agents) |
  710. | Claude Code CLI | [Example](examples/python/code/claude_code_example.py) | [📖](https://praison.ai/docs/code/claude-code) |
  711. | Gemini CLI | [Example](examples/python/code/gemini_cli_example.py) | [📖](https://praison.ai/docs/code/gemini-cli) |
  712. | Codex CLI | [Example](examples/python/code/codex_cli_example.py) | [📖](https://praison.ai/docs/code/codex-cli) |
  713. | Cursor CLI | [Example](examples/python/code/cursor_cli_example.py) | [📖](https://praison.ai/docs/code/cursor-cli) |
  714. </details>
  715. <details>
  716. <summary><strong>🧠 Memory & Knowledge</strong></summary>
  717. | Feature | Code | Docs |
  718. |---------|:----:|:----:|
  719. | Memory (Short & Long Term) | [Example](examples/python/general/memory_example.py) | [📖](https://praison.ai/docs/concepts/memory) |
  720. | File-Based Memory | [Example](examples/python/general/memory_example.py) | [📖](https://praison.ai/docs/concepts/memory) |
  721. | Claude Memory Tool | [Example](examples/python/memory/claude_memory_example.py) | [📖](https://praison.ai/docs/features/claude-memory-tool) |
  722. | Add Custom Knowledge | [Example](examples/python/concepts/knowledge-agents.py) | [📖](https://praison.ai/docs/concepts/knowledge) |
  723. | RAG Agents | [Example](examples/python/concepts/rag-agents.py) | [📖](https://praison.ai/docs/concepts/rag) |
  724. | Chat with PDF Agents | [Example](examples/python/concepts/chat-with-pdf.py) | [📖](https://praison.ai/docs/features/chat-with-pdf) |
  725. | Data Readers (PDF, DOCX, etc.) | [CLI](https://praison.ai/docs/cli/knowledge) | [📖](https://praison.ai/docs/features/chunking-strategies) |
  726. | Vector Store Selection | [CLI](https://praison.ai/docs/cli/knowledge) | [📖](https://praison.ai/docs/features/knowledge-backends) |
  727. | Retrieval Strategies | [CLI](https://praison.ai/docs/cli/knowledge) | [📖](https://praison.ai/docs/features/retrieval-strategies) |
  728. | Rerankers | [CLI](https://praison.ai/docs/cli/knowledge) | [📖](https://praison.ai/docs/features/smart-retrieval) |
  729. | Index Types (Vector/Keyword/Hybrid) | [CLI](https://praison.ai/docs/cli/knowledge) | [📖](https://praison.ai/docs/features/incremental-indexing) |
  730. | Query Engines (Sub-Question, etc.) | [CLI](https://praison.ai/docs/cli/knowledge) | [📖](https://praison.ai/docs/features/retrieval) |
  731. </details>
  732. <details>
  733. <summary><strong>🔬 Research & Intelligence</strong></summary>
  734. | Feature | Code | Docs |
  735. |---------|:----:|:----:|
  736. | Deep Research Agents | [Example](examples/python/agents/research-agent.py) | [📖](https://praison.ai/docs/agents/deep-research) |
  737. | Query Rewriter Agent | [Example](examples/python/agents/query-rewriter-agent.py) | [📖](https://praison.ai/docs/agents/query-rewriter) |
  738. | Native Web Search | [Example](examples/python/agents/websearch-agent.py) | [📖](https://praison.ai/docs/agents/websearch) |
  739. | Built-in Search Tools | [Example](examples/python/agents/websearch-agent.py) | [📖](https://praison.ai/docs/tools/tavily) |
  740. | Unified Web Search | [Example](examples/python/web_search_example.py) | [📖](https://praison.ai/docs/tools/web-search) |
  741. | Web Fetch (Anthropic) | [Example](examples/python/agents/web-fetch-agent.py) | [📖](https://praison.ai/docs/features/model-capabilities) |
  742. </details>
  743. <details>
  744. <summary><strong>📋 Planning & Execution</strong></summary>
  745. | Feature | Code | Docs |
  746. |---------|:----:|:----:|
  747. | Planning Mode | [Example](examples/python/agents/planning-agent.py) | [📖](https://praison.ai/docs/concepts/planning) |
  748. | Planning Tools | [Example](examples/python/agents/planning-agent.py) | [📖](https://praison.ai/docs/concepts/planning) |
  749. | Planning Reasoning | [Example](examples/python/agents/planning-agent.py) | [📖](https://praison.ai/docs/concepts/planning) |
  750. | Prompt Chaining | [Example](examples/python/general/prompt_chaining.py) | [📖](https://praison.ai/docs/features/promptchaining) |
  751. | Evaluator Optimiser | [Example](examples/python/general/evaluator-optimiser.py) | [📖](https://praison.ai/docs/concepts/evaluation) |
  752. | Orchestrator Workers | [Example](examples/python/general/orchestrator-workers.py) | [📖](https://praison.ai/docs/concepts/orchestration) |
  753. </details>
  754. <details>
  755. <summary><strong>👥 Specialized Agents</strong></summary>
  756. | Feature | Code | Docs |
  757. |---------|:----:|:----:|
  758. | Data Analyst Agent | [Example](examples/python/agents/data-analyst-agent.py) | [📖](https://praison.ai/docs/agents/data-analyst) |
  759. | Finance Agent | [Example](examples/python/agents/finance-agent.py) | [📖](https://praison.ai/docs/agents/finance) |
  760. | Shopping Agent | [Example](examples/python/agents/shopping-agent.py) | [📖](https://praison.ai/docs/agents/shopping) |
  761. | Recommendation Agent | [Example](examples/python/agents/recommendation-agent.py) | [📖](https://praison.ai/docs/agents/recommendation) |
  762. | Wikipedia Agent | [Example](examples/python/agents/wikipedia-agent.py) | [📖](https://praison.ai/docs/agents/wikipedia) |
  763. | Programming Agent | [Example](examples/python/agents/programming-agent.py) | [📖](https://praison.ai/docs/agents/programming) |
  764. | Math Agents | [Example](examples/python/agents/math-agent.py) | [📖](https://praison.ai/docs/features/mathagent) |
  765. | Markdown Agent | [Example](examples/python/agents/markdown-agent.py) | [📖](https://praison.ai/docs/agents/markdown) |
  766. | Prompt Expander Agent | [Example](examples/python/agents/prompt-expander-agent.py) | [📖](https://praison.ai/docs/agents/prompt-expander) |
  767. </details>
  768. <details>
  769. <summary><strong>🎨 Media & Multimodal</strong></summary>
  770. | Feature | Code | Docs |
  771. |---------|:----:|:----:|
  772. | Image Generation Agent | [Example](examples/python/image/image-agent.py) | [📖](https://praison.ai/docs/features/image-generation) |
  773. | Image to Text Agent | [Example](examples/python/agents/image-to-text-agent.py) | [📖](https://praison.ai/docs/agents/image-to-text) |
  774. | Video Agent | [Example](examples/python/agents/video-agent.py) | [📖](https://praison.ai/docs/agents/video) |
  775. | Camera Integration | [Example](examples/python/camera/) | [📖](https://praison.ai/docs/features/camera-integration) |
  776. </details>
  777. <details>
  778. <summary><strong>🔌 Protocols & Integration</strong></summary>
  779. | Feature | Code | Docs |
  780. |---------|:----:|:----:|
  781. | MCP Transports | [Example](examples/python/mcp/mcp-transports-overview.py) | [📖](https://praison.ai/docs/mcp/transports) |
  782. | WebSocket MCP | [Example](examples/python/mcp/websocket-mcp.py) | [📖](https://praison.ai/docs/mcp/sse-transport) |
  783. | MCP Security | [Example](examples/python/mcp/mcp-security.py) | [📖](https://praison.ai/docs/mcp/transports) |
  784. | MCP Resumability | [Example](examples/python/mcp/mcp-resumability.py) | [📖](https://praison.ai/docs/mcp/sse-transport) |
  785. | MCP Config Management | [Docs](https://praison.ai/docs/cli/mcp) | [📖](https://praison.ai/docs/cli/mcp) |
  786. | LangChain Integrated Agents | [Example](examples/python/general/langchain_example.py) | [📖](https://praison.ai/docs/features/langchain) |
  787. </details>
  788. <details>
  789. <summary><strong>🛡️ Safety & Control</strong></summary>
  790. | Feature | Code | Docs |
  791. |---------|:----:|:----:|
  792. | Guardrails | [Example](examples/python/guardrails/comprehensive-guardrails-example.py) | [📖](https://praison.ai/docs/concepts/guardrails) |
  793. | Human Approval | [Example](examples/python/general/human_approval_example.py) | [📖](https://praison.ai/docs/concepts/approval) |
  794. | Rules & Instructions | [Docs](https://praison.ai/docs/features/rules) | [📖](https://praison.ai/docs/features/rules) |
  795. </details>
  796. <details>
  797. <summary><strong>⚙️ Advanced Features</strong></summary>
  798. | Feature | Code | Docs |
  799. |---------|:----:|:----:|
  800. | Async & Parallel Processing | [Example](examples/python/general/async_example.py) | [📖](https://praison.ai/docs/features/async) |
  801. | Parallelisation | [Example](examples/python/general/parallelisation.py) | [📖](https://praison.ai/docs/features/parallelisation) |
  802. | Repetitive Agents | [Example](examples/python/concepts/repetitive-agents.py) | [📖](https://praison.ai/docs/features/repetitive) |
  803. | Agent Handoffs | [Example](examples/python/handoff/handoff_basic.py) | [📖](https://praison.ai/docs/concepts/handoffs) |
  804. | Stateful Agents | [Example](examples/python/stateful/workflow-state-example.py) | [📖](https://praison.ai/docs/features/stateful-agents) |
  805. | Autonomous Workflow | [Example](examples/python/general/autonomous-agent.py) | [📖](https://praison.ai/docs/concepts/autonomy) |
  806. | Structured Output Agents | [Example](examples/python/general/structured_agents_example.py) | [📖](https://praison.ai/docs/features/structured) |
  807. | Model Router | [Example](examples/python/agents/router-agent-cost-optimization.py) | [📖](https://praison.ai/docs/features/model-router) |
  808. | Prompt Caching | [Example](examples/python/agents/prompt-caching-agent.py) | [📖](https://praison.ai/docs/features/model-capabilities) |
  809. | Fast Context | [Example](examples/context/00_agent_fast_context_basic.py) | [📖](https://praison.ai/docs/features/fast-context) |
  810. </details>
  811. <details>
  812. <summary><strong>🛠️ Tools & Configuration</strong></summary>
  813. | Feature | Code | Docs |
  814. |---------|:----:|:----:|
  815. | 100+ Custom Tools | [Example](examples/python/general/tools_example.py) | [📖](https://praison.ai/docs/tools/tools) |
  816. | YAML Configuration | [Example](examples/cookbooks/yaml/secondary_market_research_agents.yaml) | [📖](https://praison.ai/docs/developers/agents-playbook) |
  817. | 100+ LLM Support | [Example](examples/python/providers/openai/openai_gpt4_example.py) | [📖](https://praison.ai/docs/models) |
  818. | Callback Agents | [Example](examples/python/general/advanced-callback-systems.py) | [📖](https://praison.ai/docs/concepts/hooks) |
  819. | Hooks | [Example](examples/python/hooks/hooks_example.py) | [📖](https://praison.ai/docs/concepts/hooks) |
  820. | Middleware System | [Example](examples/middleware/basic_middleware.py) | [📖](https://praison.ai/docs/features/middleware) |
  821. | Configurable Model | [Example](examples/middleware/configurable_model.py) | [📖](https://praison.ai/docs/features/configurable-model) |
  822. | Rate Limiter | [Example](examples/middleware/rate_limiter.py) | [📖](https://praison.ai/docs/features/rate-limiter) |
  823. | Injected Tool State | [Example](examples/middleware/injected_state.py) | [📖](https://praison.ai/docs/features/injected-state) |
  824. | Shadow Git Checkpoints | [Example](examples/checkpoints/basic_checkpoint.py) | [📖](https://praison.ai/docs/features/checkpoints) |
  825. | Background Tasks | [Example](examples/background/basic_background.py) | [📖](https://praison.ai/docs/features/background-tasks) |
  826. | Policy Engine | [Example](examples/policy/basic_policy.py) | [📖](https://praison.ai/docs/features/policy-engine) |
  827. | Thinking Budgets | [Example](examples/thinking/basic_thinking.py) | [📖](https://praison.ai/docs/features/thinking-budgets) |
  828. | Output Styles | [Example](examples/output/basic_output.py) | [📖](https://praison.ai/docs/features/output-styles) |
  829. | Context Compaction | [Example](examples/compaction/basic_compaction.py) | [📖](https://praison.ai/docs/features/context-compaction) |
  830. </details>
  831. <details>
  832. <summary><strong>📊 Monitoring & Management</strong></summary>
  833. | Feature | Code | Docs |
  834. |---------|:----:|:----:|
  835. | Sessions Management | [Example](examples/python/sessions/comprehensive-session-management.py) | [📖](https://praison.ai/docs/concepts/session-management) |
  836. | Auto-Save Sessions | [Docs](https://praison.ai/docs/cli/session) | [📖](https://praison.ai/docs/cli/session) |
  837. | History in Context | [Docs](https://praison.ai/docs/cli/session) | [📖](https://praison.ai/docs/cli/session) |
  838. | Telemetry | [Example](examples/python/telemetry/production-telemetry-example.py) | [📖](https://praison.ai/docs/features/telemetry) |
  839. | Langfuse Tracing | [Docs](https://praison.ai/docs/observability/langfuse) | [📖](https://praison.ai/docs/observability/langfuse) |
  840. | Project Docs (.praison/docs/) | [Docs](https://praison.ai/docs/cli/docs) | [📖](https://praison.ai/docs/cli/docs) |
  841. | AI Commit Messages | [Docs](https://praison.ai/docs/cli/commit) | [📖](https://praison.ai/docs/cli/commit) |
  842. | @Mentions in Prompts | [Docs](https://praison.ai/docs/cli/mentions) | [📖](https://praison.ai/docs/cli/mentions) |
  843. </details>
  844. <details>
  845. <summary><strong>🖥️ CLI Features</strong></summary>
  846. | Feature | Code | Docs |
  847. |---------|:----:|:----:|
  848. | Slash Commands | [Example](examples/python/cli/slash_commands_example.py) | [📖](https://praison.ai/docs/cli/slash-commands) |
  849. | Autonomy Modes | [Example](examples/python/cli/autonomy_modes_example.py) | [📖](https://praison.ai/docs/cli/autonomy-modes) |
  850. | Cost Tracking | [Example](examples/python/cli/cost_tracking_example.py) | [📖](https://praison.ai/docs/cli/cost-tracking) |
  851. | Repository Map | [Example](examples/python/cli/repo_map_example.py) | [📖](https://praison.ai/docs/cli/repo-map) |
  852. | Interactive TUI | [Example](examples/python/cli/interactive_tui_example.py) | [📖](https://praison.ai/docs/cli/interactive-tui) |
  853. | Git Integration | [Example](examples/python/cli/git_integration_example.py) | [📖](https://praison.ai/docs/cli/git-integration) |
  854. | Sandbox Execution | [Example](examples/python/cli/sandbox_execution_example.py) | [📖](https://praison.ai/docs/cli/sandbox-execution) |
  855. | CLI Compare | [Example](examples/compare/cli_compare_basic.py) | [📖](https://praison.ai/docs/cli/compare) |
  856. | Profile/Benchmark | [Docs](https://praison.ai/docs/cli/profile) | [📖](https://praison.ai/docs/cli/profile) |
  857. | Auto Mode | [Docs](https://praison.ai/docs/cli/auto) | [📖](https://praison.ai/docs/cli/auto) |
  858. | Init | [Docs](https://praison.ai/docs/cli/init) | [📖](https://praison.ai/docs/cli/init) |
  859. | File Input | [Docs](https://praison.ai/docs/cli/file-input) | [📖](https://praison.ai/docs/cli/file-input) |
  860. | Final Agent | [Docs](https://praison.ai/docs/cli/final-agent) | [📖](https://praison.ai/docs/cli/final-agent) |
  861. | Max Tokens | [Docs](https://praison.ai/docs/cli/max-tokens) | [📖](https://praison.ai/docs/cli/max-tokens) |
  862. </details>
  863. <details>
  864. <summary><strong>🧪 Evaluation</strong></summary>
  865. | Feature | Code | Docs |
  866. |---------|:----:|:----:|
  867. | Accuracy Evaluation | [Example](examples/eval/accuracy_example.py) | [📖](https://praison.ai/docs/cli/eval) |
  868. | Performance Evaluation | [Example](examples/eval/performance_example.py) | [📖](https://praison.ai/docs/cli/eval) |
  869. | Reliability Evaluation | [Example](examples/eval/reliability_example.py) | [📖](https://praison.ai/docs/cli/eval) |
  870. | Criteria Evaluation | [Example](examples/eval/criteria_example.py) | [📖](https://praison.ai/docs/cli/eval) |
  871. </details>
  872. <details>
  873. <summary><strong>🎯 Agent Skills</strong></summary>
  874. | Feature | Code | Docs |
  875. |---------|:----:|:----:|
  876. | Skills Management | [Example](examples/skills/basic_skill_usage.py) | [📖](https://praison.ai/docs/concepts/skills) |
  877. | Custom Skills | [Example](examples/skills/custom_skill_example.py) | [📖](https://praison.ai/docs/concepts/skills) |
  878. </details>
  879. <details>
  880. <summary><strong>⏰ 24/7 Scheduling</strong></summary>
  881. | Feature | Code | Docs |
  882. |---------|:----:|:----:|
  883. | Agent Scheduler | [Example](examples/python/scheduled_agents/news_checker_live.py) | [📖](https://praison.ai/docs/cli/scheduler) |
  884. </details>
  885. </details>
  886. ---
  887. ## 💻 Using JavaScript Code
  888. ```bash
  889. npm install praisonai
  890. export OPENAI_API_KEY=xxxxxxxxxxxxxxxxxxxxxx
  891. ```
  892. ```javascript
  893. const { Agent } = require('praisonai');
  894. const agent = new Agent({ instructions: 'You are a helpful AI assistant' });
  895. agent.start('Write a movie script about a robot in Mars');
  896. ```
  897. ---
  898. ## ⚡ Performance
  899. PraisonAI keeps framework overhead microsecond-class: creating an `Agent` (one tool, no LLM call) averaged **46 μs** in our 2026-10-15 run, so per-agent setup is negligible next to model latency in multi-agent workloads.
  900. | Performance Metric | PraisonAI | Measured |
  901. |--------------------|-----------|----------|
  902. | Avg Agent instantiation (1 tool, 1000 iterations) | **46 μs** | 2026-10-15 · praisonaiagents 1.7.12 |
  903. Methodology, comparisons and how to reproduce: [benchmarks/BENCHMARK_RESULTS.md](https://github.com/MervinPraison/PraisonAI/blob/main/src/praisonai-agents/benchmarks/BENCHMARK_RESULTS.md) · [docs](https://praison.ai/docs/developers/benchmarks)
  904. ---
  905. ---
  906. ## ⭐ Star History
  907. [![Star History Chart](https://api.star-history.com/svg?repos=MervinPraison/PraisonAI&type=Date)](https://praison.ai/docs)
  908. ---
  909. <p align="center">
  910. <img src=".github/images/agentflow.gif" alt="PraisonAI AgentFlow" width="800" />
  911. </p>
  912. <p align="center">
  913. <img src=".github/images/latest_ai_news_and_crawl_each_url_to_find_info.gif" alt="PraisonAI command execution" width="800" />
  914. </p>
  915. \* `export TAVILY_API_KEY=xxxxx`
  916. ## 🔍 Langfuse Tracing
  917. ```bash
  918. pip install "praisonai[langfuse]"
  919. praisonai langfuse
  920. ```
  921. <p align="center">
  922. <img src=".github/images/langfuse.png" alt="PraisonAI Langfuse Tracing" width="800" />
  923. </p>
  924. ---
  925. ## 🎓 Video Tutorials
  926. Learn PraisonAI through our comprehensive video series:
  927. <details>
  928. <summary><strong>View all 22 video tutorials</strong></summary>
  929. | Topic | Video |
  930. |-------|--------|
  931. | AI Agents with Self Reflection | [![Self Reflection](https://img.youtube.com/vi/vLXobEN2Vc8/0.jpg)](https://www.youtube.com/watch?v=vLXobEN2Vc8) |
  932. | Reasoning Data Generating Agent | [![Reasoning Data](https://img.youtube.com/vi/fUT332Y2zA8/0.jpg)](https://www.youtube.com/watch?v=fUT332Y2zA8) |
  933. | AI Agents with Reasoning | [![Reasoning](https://img.youtube.com/vi/KNDVWGN3TpM/0.jpg)](https://www.youtube.com/watch?v=KNDVWGN3TpM) |
  934. | Multimodal AI Agents | [![Multimodal](https://img.youtube.com/vi/hjAWmUT1qqY/0.jpg)](https://www.youtube.com/watch?v=hjAWmUT1qqY) |
  935. | AI Agents Workflow | [![Workflow](https://img.youtube.com/vi/yWTH44QPl2A/0.jpg)](https://www.youtube.com/watch?v=yWTH44QPl2A) |
  936. | Async AI Agents | [![Async](https://img.youtube.com/vi/VhVQfgo00LE/0.jpg)](https://www.youtube.com/watch?v=VhVQfgo00LE) |
  937. | Mini AI Agents | [![Mini](https://img.youtube.com/vi/OkvYp5aAGSg/0.jpg)](https://www.youtube.com/watch?v=OkvYp5aAGSg) |
  938. | AI Agents with Memory | [![Memory](https://img.youtube.com/vi/1hVfVxvPnnQ/0.jpg)](https://www.youtube.com/watch?v=1hVfVxvPnnQ) |
  939. | Repetitive Agents | [![Repetitive](https://img.youtube.com/vi/dAYGxsjDOPg/0.jpg)](https://www.youtube.com/watch?v=dAYGxsjDOPg) |
  940. | Introduction | [![Introduction](https://img.youtube.com/vi/Fn1lQjC0GO0/0.jpg)](https://www.youtube.com/watch?v=Fn1lQjC0GO0) |
  941. | Tools Overview | [![Tools Overview](https://img.youtube.com/vi/XaQRgRpV7jo/0.jpg)](https://www.youtube.com/watch?v=XaQRgRpV7jo) |
  942. | Custom Tools | [![Custom Tools](https://img.youtube.com/vi/JSU2Rndh06c/0.jpg)](https://www.youtube.com/watch?v=JSU2Rndh06c) |
  943. | Firecrawl Integration | [![Firecrawl](https://img.youtube.com/vi/UoqUDcLcOYo/0.jpg)](https://www.youtube.com/watch?v=UoqUDcLcOYo) |
  944. | User Interface | [![UI](https://img.youtube.com/vi/tg-ZjNl3OCg/0.jpg)](https://www.youtube.com/watch?v=tg-ZjNl3OCg) |
  945. | Crawl4AI Integration | [![Crawl4AI](https://img.youtube.com/vi/KAvuVUh0XU8/0.jpg)](https://www.youtube.com/watch?v=KAvuVUh0XU8) |
  946. | Chat Interface | [![Chat](https://img.youtube.com/vi/sw3uDqn2h1Y/0.jpg)](https://www.youtube.com/watch?v=sw3uDqn2h1Y) |
  947. | Code Interface | [![Code](https://img.youtube.com/vi/_5jQayO-MQY/0.jpg)](https://www.youtube.com/watch?v=_5jQayO-MQY) |
  948. | Mem0 Integration | [![Mem0](https://img.youtube.com/vi/KIGSgRxf1cY/0.jpg)](https://www.youtube.com/watch?v=KIGSgRxf1cY) |
  949. | Training | [![Training](https://img.youtube.com/vi/aLawE8kwCrI/0.jpg)](https://www.youtube.com/watch?v=aLawE8kwCrI) |
  950. | Realtime Voice Interface | [![Realtime](https://img.youtube.com/vi/frRHfevTCSw/0.jpg)](https://www.youtube.com/watch?v=frRHfevTCSw) |
  951. | Call Interface | [![Call](https://img.youtube.com/vi/m1cwrUG2iAk/0.jpg)](https://www.youtube.com/watch?v=m1cwrUG2iAk) |
  952. | Reasoning Extract Agents | [![Reasoning Extract](https://img.youtube.com/vi/2PPamsADjJA/0.jpg)](https://www.youtube.com/watch?v=2PPamsADjJA) |
  953. </details>
  954. ---
  955. ## 👥 Contributing
  956. We welcome contributions! Fork the repo, create a branch, and submit a PR → [Contributing Guide](https://github.com/MervinPraison/PraisonAI/blob/main/CONTRIBUTING.md).
  957. ---
  958. ## ❓ FAQ & Troubleshooting
  959. <details>
  960. <summary><strong>Show FAQ and troubleshooting</strong></summary>
  961. <details>
  962. <summary><strong>ModuleNotFoundError: No module named 'praisonaiagents'</strong></summary>
  963. Install the package:
  964. ```bash
  965. pip install praisonaiagents
  966. ```
  967. </details>
  968. <details>
  969. <summary><strong>API key not found / Authentication error</strong></summary>
  970. Ensure your API key is set:
  971. ```bash
  972. export OPENAI_API_KEY=your_key_here
  973. ```
  974. For other providers, see [Models docs](https://praison.ai/docs/models).
  975. </details>
  976. <details>
  977. <summary><strong>How do I use a local model (Ollama)?</strong></summary>
  978. ```bash
  979. # Start Ollama and pull a model
  980. ollama serve
  981. ollama pull llama3.2
  982. ```
  983. ```python
  984. from praisonaiagents import Agent
  985. agent = Agent(instructions="You are a helpful assistant", llm="ollama/llama3.2")
  986. agent.start("Why is the sky blue?")
  987. ```
  988. The `ollama/` prefix is what selects Ollama's handling — tool-call repair,
  989. tool-result formatting and the streaming rules small local models need. Or set
  990. it by environment instead:
  991. ```bash
  992. export OPENAI_MODEL_NAME=ollama/llama3.2
  993. ```
  994. Setting only `OPENAI_BASE_URL` is not enough: with no model named, the OpenAI
  995. default (`gpt-4o-mini`) is sent to Ollama, which answers
  996. `404 model 'gpt-4o-mini' not found`. Always name the model.
  997. Point at a non-default host with `base_url=` or `OLLAMA_HOST`:
  998. ```python
  999. agent = Agent(instructions="...", llm="ollama/llama3.2", base_url="http://192.168.1.10:11434")
  1000. ```
  1001. The same shape works for other local runtimes — `lm_studio/`, `vllm/` and
  1002. `hosted_vllm/` with their server's `base_url`.
  1003. See [Models docs](https://praison.ai/docs/models) for more details.
  1004. </details>
  1005. <details>
  1006. <summary><strong>How do I persist conversations to a database?</strong></summary>
  1007. Use the `db` parameter:
  1008. ```python
  1009. from praisonaiagents import Agent, db
  1010. agent = Agent(
  1011. name="Assistant",
  1012. memory={
  1013. "db": db(database_url="postgresql://localhost/mydb"),
  1014. "session_id": "my-session",
  1015. },
  1016. )
  1017. ```
  1018. See [Persistence docs](https://praison.ai/docs/databases/overview) for supported databases.
  1019. </details>
  1020. <details>
  1021. <summary><strong>How do I enable agent memory?</strong></summary>
  1022. ```python
  1023. from praisonaiagents import Agent
  1024. agent = Agent(
  1025. name="Assistant",
  1026. # Enables file-based memory (no extra deps!)
  1027. memory={"user_id": "user123"},
  1028. )
  1029. ```
  1030. See [Memory docs](https://praison.ai/docs/concepts/memory) for more options.
  1031. </details>
  1032. <details>
  1033. <summary><strong>How do I run multiple agents together?</strong></summary>
  1034. ```python
  1035. from praisonaiagents import Agent, Agents
  1036. agent1 = Agent(instructions="Research topics")
  1037. agent2 = Agent(instructions="Summarize findings")
  1038. agents = Agents(agents=[agent1, agent2])
  1039. agents.start()
  1040. ```
  1041. See [Agents docs](https://praison.ai/docs/concepts/agents) for more examples.
  1042. </details>
  1043. <details>
  1044. <summary><strong>How do I use MCP tools?</strong></summary>
  1045. ```python
  1046. from praisonaiagents import Agent, MCP
  1047. agent = Agent(
  1048. tools=MCP("npx @modelcontextprotocol/server-memory")
  1049. )
  1050. ```
  1051. See [MCP docs](https://praison.ai/docs/mcp/transports) for all transport options.
  1052. </details>
  1053. </details>
  1054. ### Getting Help
  1055. - 📚 [Full Documentation](https://praison.ai/docs)
  1056. - 🐛 [Report Issues](https://github.com/MervinPraison/PraisonAI/issues)
  1057. - 💬 [Discussions](https://github.com/MervinPraison/PraisonAI/discussions)
  1058. ---
  1059. <div align="center">
  1060. <p><strong>Made with ❤️ by the PraisonAI Team</strong></p>
  1061. <p>
  1062. <a href="https://praison.ai/docs">📚 Documentation</a> •
  1063. <a href="https://github.com/MervinPraison/PraisonAI">GitHub</a> •
  1064. <a href="https://youtube.com/@MervinPraison">▶️ YouTube</a> •
  1065. <a href="https://x.com/MervinPraison">𝕏 X</a> •
  1066. <a href="https://linkedin.com/in/mervinpraison">💼 LinkedIn</a>
  1067. </p>
  1068. </div>

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.