Skip to content
Back to search

better-chatbot

Formerly Better Chatbot. Navigator is an open-source AI workspace for agents, MCP and workflow automation.

Safety scan

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

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/keinsaasforever--better-chatbot/badge.svg)](/listings/keinsaasforever--better-chatbot)

HTML

<a href="/listings/keinsaasforever--better-chatbot"><img src="/api/listings/keinsaasforever--better-chatbot/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.

> **Get your app running in minutes! No installation or payment required.**

You only need **one AI Provider API Key** (OpenAI, Claude, Gemini, etc.). Everything else runs on free tiers - database, file storage, and hosting.

πŸ‘‰ **[Click this guide to deploy your site with just a few clicks](docs/tips-guides/vercel.md)**

---

**Keinsaas Navigator** - A better open-source AI chatbot for individuals and teams, inspired by ChatGPT, Claude, Grok, and Gemini.

β€’ **Multi-AI Support** - Integrates all major LLMs: OpenAI, Anthropic, Google, xAI, Ollama, and more β€’ **Powerful Tools** - MCP protocol, web search, JS/Python code execution, data visualization β€’ **Image Generation** - Create and edit images with AI models (OpenAI, Google Gemini, xAI) β€’ **Automation** - Custom agents, visual workflows, artifact generation β€’ **Collaboration** - Share agents, workflows, and MCP configurations with your team β€’ **Voice Assistant** - Realtime voice chat with full MCP tool integration β€’ **Intuitive UX** - Instantly invoke any feature with `@mention` β€’ **Quick Start** - Deploy free with Vercel Deploy button

Built with Vercel AI SDK and Next.js, combining the best features of leading AI services into one platform.

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 Β· README Β· 363 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

This file has changed since it was scanned, so this is not exactly the text that was scanned, and no line is marked. The findings are listed with the safety scan above.

  1. # keinsaas Navigator – Community Edition
  2. <div align="center">
  3. <img src="docs/assets/navigator-banner.webp" alt="keinsaas Navigator: your operating system for the AI era. Powered by agents, skills, MCP and workflows." width="100%" />
  4. <sub>formerly <b>better-chatbot</b></sub>
  5. **An open-source AI chat and agent platform you can host yourself.**
  6. Talk to any model, attach your files, connect MCP tool servers, build workflows,
  7. and give agents a place to work β€” on infrastructure you control.
  8. [![License: MIT](https://img.shields.io/badge/License-MIT-blue.svg)](LICENSE)
  9. [![Discord](https://img.shields.io/badge/Discord-join-5865F2?logo=discord&logoColor=white)](https://discord.gg/9dwdAYKBPp)
  10. </div>
  11. ---
  12. ## Deploy to Vercel
  13. [![Deploy with Vercel](https://vercel.com/button)](https://vercel.com/new/clone?repository-url=https%3A%2F%2Fgithub.com%2Fkeinsaasforever%2Fbetter-chatbot&project-name=better-chatbot&repository-name=better-chatbot&root-directory=apps%2Fweb&env=BETTER_AUTH_SECRET%2CAI_GATEWAY_API_KEY%2CCOMPOSIO_API_KEY%2CTENSORX_API_KEY%2CASSEMBLYAI_API_KEY%2CEXA_API_KEY%2CE2B_API_KEY%2CINNGEST_EVENT_KEY%2CINNGEST_SIGNING_KEY&envDefaults=%7B%22AI_GATEWAY_API_KEY%22%3A%22required%22%2C%22COMPOSIO_API_KEY%22%3A%22optional%22%2C%22TENSORX_API_KEY%22%3A%22optional%22%2C%22ASSEMBLYAI_API_KEY%22%3A%22optional%22%2C%22EXA_API_KEY%22%3A%22optional%22%2C%22E2B_API_KEY%22%3A%22optional%22%2C%22INNGEST_EVENT_KEY%22%3A%22optional%22%2C%22INNGEST_SIGNING_KEY%22%3A%22optional%22%7D&envDescription=Required%3A%20BETTER_AUTH_SECRET%20%28any%20long%20random%20string%2C%20e.g.%20openssl%20rand%20-base64%2032%29%20and%20AI_GATEWAY_API_KEY%20%28vercel.com%2Fai-gateway%3B%20serves%20every%20model%2C%20including%20realtime%20voice%29.%20Optional%2C%20leave%20at%20optional%20to%20switch%20the%20feature%20off%3A%20COMPOSIO_API_KEY%20%28connected%20apps%29%2C%20TENSORX_API_KEY%20%28EU-hosted%20models%29%2C%20ASSEMBLYAI_API_KEY%20%28voice%20transcription%29%2C%20EXA_API_KEY%20%28web%20search%29%2C%20E2B_API_KEY%20%28code%20sandboxes%29%2C%20INNGEST_EVENT_KEY%20and%20INNGEST_SIGNING_KEY%20%28scheduled%20agents%29.%20Add%20any%20of%20them%20later%20under%20Settings%2C%20Environment%20Variables.&envLink=https%3A%2F%2Fgithub.com%2Fkeinsaasforever%2Fbetter-chatbot%2Fblob%2Fmain%2Fdocs%2Fself-hosting.md%23vercel&stores=%5B%7B%22type%22%3A%22integration%22%2C%22integrationSlug%22%3A%22supabase%22%2C%22productSlug%22%3A%22supabase%22%2C%22protocol%22%3A%22storage%22%7D%5D)
  14. Two required values, a few optional ones. The button clones the repository,
  15. creates a free Supabase project for the database and file uploads, and asks
  16. for:
  17. | Variable | |
  18. | --- | --- |
  19. | `BETTER_AUTH_SECRET` | **Required.** Any long random string (`openssl rand -base64 32`). |
  20. | `AI_GATEWAY_API_KEY` | **Required.** From [vercel.com/ai-gateway](https://vercel.com/ai-gateway). Serves every model in the catalog, including realtime voice. |
  21. | `COMPOSIO_API_KEY` | Optional. Lets users connect Gmail, Slack, GitHub and more. [Free key](https://platform.composio.dev) |
  22. | `TENSORX_API_KEY` | Optional. EU-hosted models from our [partner](#partners) TensorX. |
  23. | `ASSEMBLYAI_API_KEY` | Optional. Voice transcription via [AssemblyAI](https://www.assemblyai.com/). |
  24. | `EXA_API_KEY` | Optional. Web search via [Exa](https://exa.ai/). |
  25. | `E2B_API_KEY` | Optional. Sandboxes that run generated code, via [E2B](https://e2b.dev/). Without it, code runs in the browser. |
  26. | `INNGEST_EVENT_KEY`, `INNGEST_SIGNING_KEY` | Optional. Scheduled agents and background jobs, via [Inngest](https://www.inngest.com/). Easiest: add the Inngest integration in Vercel, which sets both and connects the app. |
  27. The optional fields are pre-filled with `optional`; leave them like that and the
  28. feature stays off. You can add any key later under Settings β†’ Environment
  29. Variables and redeploy. The database schema and the storage buckets are
  30. created during the build. Open the deployment and create an account; the first
  31. one becomes the administrator.
  32. Free Supabase projects pause after a week without use. If your app stops
  33. loading after a quiet week, open the project in the Supabase dashboard and
  34. click **Restore**.
  35. **Updates come to you.** After your first sign-in, the app asks once to
  36. **Turn on** automatic updates: it opens GitHub with the update check ready,
  37. and you click **Commit changes**. From then on your copy checks every week.
  38. When there is a new version, GitHub emails you a short guide: try the preview
  39. Vercel builds, then click **Merge**. Nothing changes in your live app until
  40. you do.
  41. Details and optional features: [docs/self-hosting.md](docs/self-hosting.md#vercel).
  42. ---
  43. ## Quick start (Docker)
  44. Three variables. That is the whole configuration to get chatting:
  45. ```bash
  46. git clone https://github.com/keinsaasforever/better-chatbot.git
  47. cd better-chatbot
  48. cp .env.example .env
  49. ```
  50. ```bash
  51. POSTGRES_URL=postgres://postgres:postgres@postgres:5432/better_chatbot
  52. BETTER_AUTH_SECRET= # npx @better-auth/cli@latest secret
  53. AI_GATEWAY_API_KEY= # https://vercel.com/ai-gateway
  54. ```
  55. ```bash
  56. docker compose -f docker/compose.yml up -d --build
  57. ```
  58. Open http://localhost:3000 and create an account. The first account becomes the
  59. administrator.
  60. **Why one model key?** Models are served through the Vercel AI Gateway, which
  61. fronts Anthropic, OpenAI, Google, Mistral, xAI, DeepSeek, Moonshot and more. You
  62. manage one key and one bill instead of seven, and you can switch models
  63. mid-conversation without touching configuration. If you would rather not use a
  64. gateway at all, point the app at Ollama, LM Studio, or any OpenAI-compatible
  65. endpoint β€” see [Models](#models).
  66. Prefer running it directly? See [docs/self-hosting.md](docs/self-hosting.md).
  67. ---
  68. ## See it in action
  69. ### Web search
  70. Ask about anything current and get an answer with its sources. Needs the optional `EXA_API_KEY`.
  71. <img src="docs/assets/demo-web-search.webp" alt="A chat that searches the web and answers with sources" width="100%" />
  72. ### Charts and tables
  73. One sentence in, a chart or an interactive table out.
  74. <img src="docs/assets/demo-chart.webp" alt="A chat that answers with a bar chart" width="100%" />
  75. ### Code execution
  76. Python and JavaScript run in a sandbox; results come back as tables and charts. Needs the optional `E2B_API_KEY`.
  77. <img src="docs/assets/demo-code.webp" alt="A chat that runs Python and shows the result as a table" width="100%" />
  78. ### Agents
  79. Mention an agent with `@` and it takes over the conversation with its own instructions and tools.
  80. <img src="docs/assets/demo-agent.webp" alt="Mentioning an agent with @ and getting a researched answer" width="100%" />
  81. ### Legal document review
  82. An example from the agent library: paste a clause or attach a contract, and the Legal Document Reviewer flags the risks and suggests changes.
  83. <img src="docs/assets/demo-legal.webp" alt="The Legal Document Reviewer agent analysing the risks of an NDA clause" width="100%" />
  84. ### Image generation
  85. Generate images right inside the conversation.
  86. <img src="docs/assets/demo-image.webp" alt="A chat that generates a watercolor image" width="100%" />
  87. ### Tools and MCP servers
  88. Connect your apps and any MCP server. The catalog of 250+ apps comes with the optional Composio integration.
  89. <img src="docs/assets/demo-mcp.webp" alt="Connected apps and the catalog of tools to connect" width="100%" />
  90. ### TensorX: EU-hosted, zero data retention
  91. Pick a model under **TensorX (0 Data Retention)** and your prompts go straight to EU infrastructure that keeps nothing. Needs the optional `TENSORX_API_KEY`; see [Partners](#partners).
  92. <img src="docs/assets/demo-tensorx.webp" alt="Choosing a TensorX model marked 0 Data Retention and chatting with it" width="100%" />
  93. ### Local models
  94. At the end of the model menu, **Add local Model** connects Ollama, LM Studio or any OpenAI-compatible server. Your data stays on your hardware.
  95. <img src="docs/assets/demo-local-models.webp" alt="Opening Add local Model at the end of the model menu and entering a server address" width="100%" />
  96. ### Voice input
  97. Click the microphone, speak, and your words are transcribed into the message box by AssemblyAI. Needs the optional `ASSEMBLYAI_API_KEY`.
  98. <img src="docs/assets/demo-voice.webp" alt="Clicking the microphone button, which turns red while recording" width="100%" />
  99. ### Desktop app (beta)
  100. The keinsaas Navigator desktop app for Windows and macOS, with local (stdio) MCP servers and native notifications. It connects to the cloud or to your own instance; see [What you get](#what-you-get).
  101. <img src="docs/assets/desktop-app.webp" alt="The keinsaas Navigator desktop app on Windows, showing the sign-in screen" width="100%" />
  102. ---
  103. ## What you get
  104. **Chat that does things.** Streaming conversations with tool calling, file
  105. attachments, image generation, and a sandboxed code runner. Mention an agent,
  106. workflow, or tool with `@` to pull it into the conversation.
  107. **Agents.** Give an agent a role, instructions, and a set of tools. Share it
  108. with your workspace or keep it private.
  109. **Skills.** Reusable capabilities an agent loads on demand, so a large toolbox
  110. does not have to sit in every prompt.
  111. **Projects with retrieval.** Group chats, documents, and tools around a piece of
  112. work. Uploaded documents are chunked, embedded, and searched with pgvector β€”
  113. no external service required. Point a project at a knowledge base if you want
  114. hosted retrieval instead.
  115. **Workflows.** Build a graph of steps β€” LLM calls, conditions, HTTP requests,
  116. code, templates β€” in a visual editor, publish it as a tool, and call it from any
  117. chat.
  118. **MCP tool servers.** Connect any Model Context Protocol server over HTTP or
  119. stdio, with OAuth support, per-tool customization and a cached tool list.
  120. Self-hosted instances can run stdio servers; the desktop app runs them locally.
  121. **Scheduled tasks.** Have an agent run on a schedule and report back.
  122. **Voice.** Realtime voice conversations through the gateway, plus transcription.
  123. **Voice device.** Pair an [M5Stack Atom Echo S3R](https://docs.m5stack.com/en/atom/Atom%20EchoS3R),
  124. press its button and talk: what you say is transcribed and handed to your
  125. agent with its tools, or turned into a new agent or a scheduled task. Pairing and device management are in
  126. the app; the device streams its audio through a small WebSocket bridge, which
  127. is not published for self-hosting yet
  128. ([#429](https://github.com/keinsaasforever/better-chatbot/issues/429)).
  129. **Desktop app (beta for self-hosting).** The keinsaas Navigator desktop app
  130. ([download at keinsaas.com](https://keinsaas.com)) adds local (stdio) MCP servers, which a server on Vercel cannot run, plus
  131. native notifications. On first launch it asks where to connect: pick **My own
  132. server** and enter your instance's address, e.g. `https://my-chatbot.vercel.app`.
  133. Connecting it to a self-hosted instance is in beta; tell us in
  134. [Discord](https://discord.gg/9dwdAYKBPp) what works and what does not. You can
  135. switch servers any time under **Change Server…** in the app menu.
  136. **Model routing.** Pick a model per message, or let the router choose from the
  137. catalog based on the task.
  138. ---
  139. ## Upgrading from 1.x
  140. This is a new major version. The layout is now a monorepo, the environment is
  141. different, and the schema has moved on considerably.
  142. **Back up your database first.** Then:
  143. ```bash
  144. pnpm install
  145. pnpm tsx scripts/upgrade/from-better-chatbot-1.ts --dry-run
  146. pnpm tsx scripts/upgrade/from-better-chatbot-1.ts --yes
  147. ```
  148. The upgrade preserves your users, chats, agents, MCP servers and workflows.
  149. Archives become projects, and the threads inside them stay linked. It is a
  150. one-way migration β€” this is why the backup matters.
  151. If you were setting per-provider API keys (`OPENAI_API_KEY`, `ANTHROPIC_API_KEY`
  152. and friends), replace them with a single `AI_GATEWAY_API_KEY`, or configure a
  153. local provider.
  154. ---
  155. ## Setup
  156. ### Vercel
  157. The [Deploy button](#deploy-to-vercel) above. Setting up an existing clone by
  158. hand is described in [docs/self-hosting.md](docs/self-hosting.md#vercel).
  159. ### Docker Compose
  160. The [quick start](#quick-start-docker) above. The bundled stack includes Postgres with
  161. pgvector. Add `--profile redis` if you want resumable chat streams.
  162. ### Local development
  163. ```bash
  164. pnpm install # writes .env from .env.example
  165. pnpm docker:pg # Postgres with pgvector on :5432
  166. pnpm db:migrate # apply the schema
  167. pnpm dev # http://localhost:3000
  168. ```
  169. Postgres **must** have the `pgvector` extension β€” document embeddings are stored
  170. as `vector(1536)`. The bundled image has it; managed Postgres usually offers it
  171. as a toggle.
  172. Useful commands:
  173. | Command | What it does |
  174. | --- | --- |
  175. | `pnpm check` | Lint, typecheck and unit tests β€” run before committing |
  176. | `pnpm test:e2e` | Playwright end-to-end tests |
  177. | `pnpm db:migrate` | Apply pending migrations |
  178. | `pnpm db:studio` | Browse the database |
  179. | `pnpm docker:redis` | Redis for resumable streams |
  180. ---
  181. ## Optional integrations
  182. Everything below is off until you configure it, except file uploads on Vercel,
  183. which the Supabase integration sets up for you. The server prints which
  184. features are disabled at startup, so you always know why something is missing.
  185. | Feature | Set | Without it |
  186. | --- | --- | --- |
  187. | File uploads | Nothing on Vercel: the Supabase integration sets it. Docker and local: `SUPABASE_URL`, `NEXT_PUBLIC_SUPABASE_URL`, `SUPABASE_SERVICE_ROLE_KEY` ([details](docs/self-hosting.md#storage)) | Attachments, avatars and generated images are disabled |
  188. | Knowledge bases | `AGENTSET_API_KEY` | Knowledge bases unavailable; project documents still search locally with pgvector |
  189. | Background jobs | `INNGEST_SIGNING_KEY`, `INNGEST_EVENT_KEY` | Scheduled tasks need your own cron (see below); the project brain runs right after each change instead of in the background |
  190. | Resumable streams | `REDIS_URL` | A stream ends when the viewer navigates away |
  191. | Code sandbox | `E2B_API_KEY` | Generated code runs in the in-browser worker only |
  192. | Web search | `EXA_API_KEY` | The web search tool is unavailable |
  193. | EU-hosted models | `TENSORX_API_KEY` | TensorX models (zero data retention) do not appear in the model picker |
  194. | Voice transcription | `ASSEMBLYAI_API_KEY` | Realtime voice still works; transcription does not |
  195. | Browser notifications | `NEXT_PUBLIC_VAPID_PUBLIC_KEY`, `VAPID_PRIVATE_KEY` | No notifications while the app is closed |
  196. | Connected apps | `COMPOSIO_API_KEY` | Users cannot connect Gmail, Slack, GitHub and other apps |
  197. | Observability | `SUPERLOG_PUBLIC_TOKEN` | Traces and logs stay local |
  198. **Running without Inngest.** Scheduled tasks and the sandbox reaper are plain
  199. HTTP endpoints. Set `SCHEDULED_TASK_SECRET` and call them from any cron:
  200. ```bash
  201. curl -X POST https://your-instance/api/scheduled-tasks/cron \
  202. -H "X-Scheduled-Task-Auth: $SCHEDULED_TASK_SECRET"
  203. ```
  204. ---
  205. ## Models
  206. The gateway is the default and covers the whole catalog. You can also:
  207. - **Run models locally.** Connect Ollama, LM Studio, or any OpenAI-compatible
  208. endpoint from the model picker. Nothing leaves your machine, and it needs no
  209. server configuration β€” each account sets up its own.
  210. - **Use [TensorX](#partners).** EU-hosted inference with zero data retention,
  211. reached directly rather than through the gateway. Set `TENSORX_API_KEY`.
  212. - **Add any OpenAI-compatible provider** with `OPENAI_COMPATIBLE_DATA`:
  213. `pnpm openai-compatiable:init && pnpm openai-compatiable:parse`.
  214. ---
  215. ## Partners
  216. **[TensorX](https://tensorx.ai/)** β€” EU-hosted model inference with zero data
  217. retention. The Community Edition talks to TensorX directly rather than through a
  218. gateway, so prompts go straight to TensorX's EU infrastructure. Create a key at
  219. [tensorx.ai](https://tensorx.ai/), set `TENSORX_API_KEY`, and its models appear
  220. in the model picker next to the gateway's. See its
  221. [trust and security](https://tensorx.ai/trust-security/) page and
  222. [DPA](https://tensorx.ai/dpa/).
  223. TensorX is optional. Without the key everything else works as before.
  224. ---
  225. ## Community
  226. - **[Discord](https://discord.gg/9dwdAYKBPp)** β€” questions, help, and what
  227. people are building
  228. - **[Issues](https://github.com/keinsaasforever/better-chatbot/issues)** β€” bugs
  229. and feature requests
  230. - **[CONTRIBUTING.md](CONTRIBUTING.md)** β€” how to work in this repository
  231. ---
  232. ## This edition, and the hosted one
  233. This repository is the Community Edition of
  234. [keinsaas Navigator](https://keinsaas.com), formerly known as better-chatbot. The
  235. repository keeps its name, so existing links, forks and automatic updates keep
  236. working. Everything above is here, in full,
  237. with no usage caps and no telemetry.
  238. What Navigator adds is what teams need and single users do not: shared team
  239. workspaces with member management and invitations, per-member AI budgets and
  240. org-wide model policy, billing, a managed workflow builder, and a Slack bot.
  241. If you are running this for yourself or a small group, you are not missing
  242. anything β€” that is the point.
  243. ---
  244. ## License
  245. MIT. See [LICENSE](LICENSE).
  246. Originally forked from [cgoinglove/better-chatbot](https://github.com/cgoinglove/better-chatbot),
  247. whose work this is built on.

Ratings and comments

Sign in to rate this skill.

Sign in

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