Skip to content
Back to search

ollaya

Run open decision models locally: pull and serve Laya, decider, NLI and GLiClass behind a TypeSafe-compatible API. Ollama for decision models.

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/ollaya-dev--ollaya/badge.svg)](/listings/ollaya-dev--ollaya)

HTML

<a href="/listings/ollaya-dev--ollaya"><img src="/api/listings/ollaya-dev--ollaya/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.

These steps run a script from the network that findsafeskills does not scan.

- **Linux** (x86_64 or arm64, glibc ≥ 2.38, e.g. Ubuntu 24.04+): `curl -fsSL https://ollaya.dev/install.sh | sh`. When an NVIDIA GPU is present (driver R525+), the installer adds the CUDA runtime: CUDA 13 for R580+, CUDA 12 for older drivers. - **macOS** (Apple silicon): the same command. - **Windows** (x64): `irm https://ollaya.dev/install.ps1 | iex` in PowerShell. When an NVIDIA GPU is present (driver R527+), the installer adds the CUDA runtime, as on Linux. - **Desktop app** for macOS, Windows and Linux: start and stop the server, download models and run them in one window. On macOS it lives in the menu bar. Get it from [ollaya.dev/download](https://ollaya.dev/download). - **Docker:** `docker run -d --gpus=all -p 11435:11435 ghcr.io/ollaya-dev/ollaya:cuda` (`:cuda12` for host drivers older than R580), or `ghcr.io/ollaya-dev/ollaya` for CPU only.

Configuration is through environment variables: `OLLAYA_HOST`, `OLLAYA_MODELS`, `OLLAYA_KEEP_ALIVE`, `OLLAYA_DEVICE`, `OLLAYA_API_KEY` and others, listed in [docs/api.md §15](docs/api.md).

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 · 255 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. <p align="center">
  2. <picture>
  3. <source media="(prefers-color-scheme: dark)" srcset="site/public/static/logo-white.svg">
  4. <img src="site/public/static/logo.svg" alt="" width="88">
  5. </picture>
  6. </p>
  7. <h1 align="center">ollaya</h1>
  8. <p align="center"><strong>The decision layer for AI agents.</strong></p>
  9. <p align="center">Ollaya runs open decision models on your own hardware, the way Ollama runs LLMs. Typed questions in, calibrated answers out, in milliseconds: routing, guardrails, triage and tool approval without an LLM call.</p>
  10. <p align="center">
  11. <a href="https://ollaya.dev">Website</a> ·
  12. <a href="https://ollaya.dev/search">Models</a> ·
  13. <a href="https://bench.ollaya.dev">Bench</a> ·
  14. <a href="https://ollaya.dev/docs">Docs</a> ·
  15. <a href="https://github.com/ollaya-dev/ollaya/releases">Releases</a> ·
  16. <a href="https://huggingface.co/ollaya-dev">Hugging Face</a>
  17. </p>
  18. <p align="center">Created and maintained by <a href="https://github.com/cobanov">Mert Cobanov</a> (<a href="https://x.com/mertcobanov">@mertcobanov</a>).</p>
  19. A decision model reads a *state* (a message, an email, a ticket, any JSON) plus typed questions
  20. (`choice`, `score`, `noul`) and returns calibrated probabilities in a single forward pass, in
  21. milliseconds. It never generates text. Ollaya pulls these models by name and serves them from a
  22. local daemon behind a self-hosted, Jev-compatible API (TypeSafe's `/v1/systemone` wire format),
  23. so the TypeSafe SDK and existing Jev clients work unchanged once they point at Ollaya and name an
  24. open model.
  25. ```sh
  26. curl -fsSL https://ollaya.dev/install.sh | sh
  27. ollaya run winnow:e4b --preset triage "Third time this year you've double-charged me. Refund it today or I'm cancelling and moving to a competitor."
  28. ```
  29. ```
  30. intent refund ███████████████░ 0.91
  31. is_urgent yes ███████████████░ 0.92
  32. frustration 2.89 / 3 very angry or using stron… ██████████████░░ 0.86
  33. refund_requested yes ████████████████ 0.99
  34. churn_risk yes ████████████████ 0.99
  35. ```
  36. Or ask your own questions: `ollaya run kev --ask "Is a refund requested?" --ask "Which team?" --choice billing --choice sales < ticket.txt`.
  37. <!-- results:recommended -->
  38. `winnow:e4b` is the recommended model: 0.724 accuracy on typed decisions (TypeSafe reports 0.738 for its Jev) and 113 ms for five questions on an RTX 5090. It is a 4B-class language model, so without an NVIDIA GPU start with `laya`: `laya:multilingual` answers five questions in 516 ms on a CPU. All models and their numbers: [ollaya.dev/search](https://ollaya.dev/search).
  39. <!-- /results:recommended -->
  40. ## A self-hosted, Jev-compatible API
  41. **What is a Jev-compatible API?** An HTTP API that takes the same requests as TypeSafe's hosted
  42. Jev (a `state` plus typed `questions` at `POST /v1/systemone`) and returns the same `answers`, so
  43. TypeSafe's SDK and any client written for Jev work against it unchanged. Ollaya's `/v1/systemone`,
  44. `/v1/decisions` and `/v1/models` are wire-identical to TypeSafe's (as `typesafe-sdk` 0.7.1 defines
  45. them), served from your own machine with open models.
  46. **Migrating from Jev.** Point the SDK at Ollaya and pick an open model:
  47. ```sh
  48. export TYPESAFE_BASE_URL=http://localhost:11435
  49. export TYPESAFE_API_KEY=local # any non-empty value
  50. export TYPESAFE_DEFAULT_MODEL=winnow:e4b # instead of "jev-latest"
  51. ```
  52. Details, including proxies and warm-up: [TypeSafe compatibility](https://ollaya.dev/docs/typesafe-compatibility).
  53. | | Ollaya | TypeSafe Jev | Ollama |
  54. |---|---|---|---|
  55. | Runs on | Your machine or server | TypeSafe's cloud | Your machine or server |
  56. | Jev API | `/v1/systemone`, `/v1/decisions`, `/v1/models` | The original | `/v1/systemone` |
  57. | Models | Open decision models from Perplexity, Cloudflare, PostHog, Bespoke Labs and others | Jev (closed) | Its own GGUF builds of nine decision models (0.40, October 2026) |
  58. | Probabilities | Each author's calibration | TypeSafe's | Raw softmax |
  59. | Your data leaves the machine | No | Yes | No |
  60. | License | Apache-2.0 | Proprietary, hosted | MIT |
  61. The [FAQ](https://ollaya.dev/docs/faq) compares Ollaya and Ollama in more detail, measured on the same GPU.
  62. ## Features
  63. - **One binary.** `ollaya serve` runs the daemon; `ollaya run`, `pull`, `list`, `ps`, `show`, `rm`,
  64. `cp`, `stop` and `create` work the way they do in Ollama. If the daemon isn't running, the CLI
  65. starts it.
  66. - **TypeSafe-compatible.** `POST /v1/systemone`, `/v1/decisions` and `GET /v1/models` are
  67. wire-identical to TypeSafe. The official SDK works unchanged when you set
  68. `TYPESAFE_BASE_URL=http://localhost:11435`.
  69. - **Native API.** `/api/decide` adds routing information and timings. `/api/pull` streams
  70. NDJSON progress, and there are `/api/tags`, `/api/show`, `/api/ps` and more. See
  71. [docs/api.md](docs/api.md).
  72. - **Weights come from their authors.** Ollaya publishes only small ONNX graphs, about 3 MB each.
  73. These graphs read the original weight files (usually `model.safetensors`) from the author's
  74. Hugging Face repository, pinned to a commit and verified by sha256. Models whose authors publish
  75. GGUF files (`winnow`, `jevk5`, `jeb`) run that file itself on llama.cpp. Ollaya never re-hosts weights.
  76. - **For agents.** `ollaya mcp` serves the models to Claude Code, Claude Desktop, Cursor and other
  77. MCP clients (`claude mcp add ollaya -- ollaya mcp`, or [one click in Cursor](https://cursor.com/install-mcp?name=ollaya&config=eyJjb21tYW5kIjoib2xsYXlhIiwiYXJncyI6WyJtY3AiXX0%3D)), and the
  78. [`ollaya-decisions` skill](skills/ollaya-decisions/SKILL.md) teaches agents when and how to use
  79. them (`npx skills add ollaya-dev/ollaya --skill ollaya-decisions`).
  80. - **Routers.** `laya` detects the script and language of each request, then answers with
  81. `laya:en` or `laya:multilingual`.
  82. - **Modelfiles.** You can bake a question set into your own model:
  83. ```
  84. FROM laya
  85. QUESTIONS ./triage.json
  86. PARAMETER precision fp32
  87. ```
  88. Then run `ollaya create triage -f Modelfile` and `ollaya run triage "…"`.
  89. - **Fast and exact.**
  90. - **Hardware:** ONNX Runtime on CPU, and CUDA on NVIDIA GPUs. GGUF models run on llama.cpp:
  91. CPU, ROCm/HIP on AMD GPUs, Vulkan on Windows (any vendor's GPU; an integrated one when you name it), CUDA on NVIDIA GPUs, and Metal on Apple silicon.
  92. - **Precision:** ONNX uses fp16 on GPU and fp32 on CPU when available; GGUF keeps its model quantization on either device.
  93. - **Accuracy:** fp32 exports give the same decision as the PyTorch reference on 100% of 2,383
  94. questions per checkpoint.
  95. ## Models
  96. | Model | What it is |
  97. |---|---|
  98. | `winnow:e4b` | **Recommended.** EldanRing's Winnow-E4B, a Gemma 4 fine-tune run from the author's Q8_0 GGUF on llama.cpp |
  99. | `credence:e4b`, `credence:e4b-calibrated` | Txoka's MiCA refinements of Winnow-E4B, Q8_0: accuracy-focused and public-calibration variants; see the measured tradeoffs in the family documentation |
  100. | `laya` | Router: picks `laya:en` or `laya:multilingual` by language |
  101. | `laya:en` | English decision model (ModernBERT-large, 421M) |
  102. | `laya:multilingual` | 100+ languages (mmBERT-base, 322M) |
  103. | `laya:typed-decisions` | Fine-tuned on the typed-decisions workflows |
  104. | `decider`, `decider:4b`, `decider:0.8b` | Mapika's Qwen3.5 decoders, 2B (the default), 4B and 0.8B |
  105. | `decider:2b-vision` | Mapika's Qwen3.5-2B vision-language decider: questions about an image (`--image`, `images` on `/api/decide`) as well as the state |
  106. | `kev`, `kev:0.8b`, `kev:9b` | Jared Palmer's Kev: a LoRA and a pointer head on Qwen3.5 (4B by default, 0.8B, 9B), calibrated |
  107. | `decision` | Decision 1.0 Eos by the vLLM Semantic Router contributors: a fine-tuned Qwen3.5-0.8B with an endpoint head, 16k-token rows |
  108. | `qwen3guard` | Qwen3Guard-Gen-0.6B safety guard with built-in questions: safe, controversial or unsafe, and the category |
  109. | `nli`, `nli:modernbert-large` | Moritz Laurer's zero-shot NLI classifiers (DeBERTa-v3-large, ModernBERT-large) |
  110. | `gliclass` | Knowledgator's instruction-following zero-shot classifier (DeBERTa-v3-large) |
  111. | `gliclass-v4`, `gliclass-v4:9b` | Knowledgator's GLiClass v4 (decoder-KV on Qwen3.5-2B and -9B): the labels follow the instructions and the state in one decoder row, a two-layer scorer with up to three reasoning steps reads them, and the model's own calibrator rescales every label; text only |
  112. | `von` | Victor Hugo Panisa's Von 1.1 (ModernBERT-large): every option scored at its own marker, 8k-token context. Archived upstream on 2026-10-19; `d1-omni` is its successor |
  113. | `winnow` | EldanRing's Winnow-12B, the larger sibling of `winnow:e4b` |
  114. | `clm` | Contrastive-LM's CLM-v0.1-8B: the Qwen3-8B encoder and two projection heads score options by similarity, with questions and options cached. Built for agent, game and tool-calling states |
  115. | `jevk5` | alibiserikbay's JevK5 v0.3, a Qwen3.5-4B fine-tune run from the author's Q8_0 GGUF on llama.cpp, up to 16 options |
  116. | `nimble`, `nimble:9b-v3` | Bespoke Labs' Nimble v2: a LoRA on Qwen3.5-9B that reads the whole request as a JSON schema and scores option codes, with the author's temperature, up to 255 options. `nimble:9b-v3` is v3 (non-commercial license) as the author's Decision Index engine serves it |
  117. | `strands-decider`, `strands-decider:e2b`, `strands-decider:e4b` | Strands Agents' Strands Decider: a LoRA and a pointer head on Qwen3.5-2B base (the default) or on Gemma 4 E2B / E4B, every option scored at the end of its own line, calibrated per question type; up to 255 options |
  118. | `pplx-decider` | Perplexity's pplx-decider-flash v1.2: Qwen3.5-4B distilled from its 27B decision models, every question in one sequence and read at its own marker, calibrated; up to 255 options |
  119. | `d3`, `d3:edge`, `d3:lite`, `d3:nano`, `d3:flash` | Decision 3.0 by the vLLM Semantic Router contributors: Qwen3.5 fine-tunes (0.6B to 9B, 4B by default) with a 255-code readout, every question its own forward pass. Images and videos are not served |
  120. | `jeb`, `jeb:4b`, `jeb:27b` | Jason Brashear's Jebadiah (AINode): LoRAs merged into Qwen3.5 (9B by default, 4B) and Qwen3.8-27B, run from the authors' GGUF on llama.cpp with their per-type temperatures |
  121. | `jeeves` | PostHog's Jeeves-9B without its reasoning chain: Qwen3.5-9B (LoRA merged) and a pointer head |
  122. | `clef` | Cloudflare's Clef-Flash: Qwen3.5-9B, fully post-trained, with a joint schema head that scores every question's options in one forward pass, with no fitted temperature |
  123. | `cygnet` | blockbrain-ai's Cygnet: frozen Gemma 4 12B IT (Q8_0 GGUF) with a letter-readout prompt and temperature 3.4 |
  124. | `snap` | logitlab's snap1-2b: MiniCPM5-2B fine-tuned on the prompt of emnlmn's snap engine, run from the author's Q8_0 GGUF on llama.cpp with snap's own prompt, up to 26 options |
  125. | `d1`, `d1:3b-vision` | Liquid AI's d1-3B, on LFM2.5-VL-3B: typed text and image decisions from the author's Q8_0 GGUF and projector, with max-pooled answer forms and temperature 1.0 |
  126. | `d1-omni` | Liquid AI's d1-omni-600M, text questions: an LFM2.5 encoder with a decision head that scores every option at its own marker, calibrated per question type and option count; 16k-token context. Its image and audio paths are not served |
  127. | `decima`, `decima:agent`, `decima:small` | A. M. Madani's Decima: multilingual encoders with a late-interaction scorer that reads every option against the state. `decima:base` is mmBERT-base (321M); `decima:agent` (2.2) is fine-tuned for coding-agent decisions and knowledge-graph judgments; `decima:small` (122M) is the smallest |
  128. | `largha` | Phocinae's Largha: Laya's decision network on mmBERT-small (144M), fine-tuned on the typed-decisions workflows; runs on the CPU, CUDA and the Apple GPU |
  129. | `arbiter` | Codekins' Arbiter v3.3 (Zyot Lab): a LoRA and a fixed 24-slot head on Gemma 3 4B IT, one forward pass per question. It answers noul, choices of up to 16 options and scores of exactly 6 levels |
  130. Browse them at [ollaya.dev/search](https://ollaya.dev/search); every tag's accuracy, speed and memory is on
  131. [Ollaya Bench](https://bench.ollaya.dev). Decoder tags ending in `-fp16` (`kev:4b-fp16`) compute in fp16 on NVIDIA
  132. GPUs: faster, with at least 99.5 % of their decisions the same as the base tag's. Laya's `-fp32` and `-fp16` tags
  133. pin the precision. The derived files of every model are also published at
  134. [huggingface.co/ollaya-dev](https://huggingface.co/ollaya-dev).
  135. Ollama 0.40 also serves decision models, nine of them, through the same TypeSafe wire format
  136. ([measured side by side](docs/measurements/2026-10-ollama-0.40.md)).
  137. [The FAQ](https://ollaya.dev/docs/faq#ollama-runs-decision-models-now-how-is-ollaya-different)
  138. compares the two projects.
  139. ## Results
  140. Every model in the library is measured on our own machines and published, with the raw data, on
  141. [Ollaya Bench](https://bench.ollaya.dev): one released runtime, each author's own weights, the same
  142. 5,880 questions for every model (Bespoke Labs' 13 public subsets and typed-decisions), and speed and
  143. memory on a GPU and a CPU.
  144. <!-- results:bench -->
  145. Edition 2026.10 ran Ollaya 0.13.0 on an RTX 5090 desktop, and 0.14.0, 0.15.0, 0.17.0, 0.18.0 and 0.19.0 added what they changed:
  146. - **Score.** Chance-corrected accuracy over 14 tasks, with 95 % intervals: `pplx-decider:flash` 73.7, `jeb:27b` 69.1 (scored without the 5 tasks it was trained on), `d3:mini` 68.0, `d3:flash` 67.8 and `jeb:9b` 66.0 (scored without the 5 tasks it was trained on). The fastest of these, `jeb:9b`, answers five questions in 111 ms on the RTX 5090.
  147. - **Calibration.** The lowest calibration errors on the public benchmark are 0.016 (`pplx-decider:flash`), 0.018 (`jeeves:9b`) and 0.020 (`jeb:9b`).
  148. - **Speed and memory.** Five-question requests through the HTTP API take 14 ms to 1.0 s on the RTX 5090 and 314 ms to 62.0 s on its Threadripper 3960X. Every model's VRAM and RAM are on its page.
  149. - **Parity.** Before a model ships, its runtime is checked question by question against the authors' code (or llama.cpp's own server, for GGUF models) on each device it runs on.
  150. Earlier runs, the comparison with Ollama on the same GPU (on the same Nimble weights the calibration error is 0.022 on Ollaya and 0.115 on Ollama) and the Windows, CUDA and Vulkan measurements are at [ollaya.dev/results](https://ollaya.dev/results).
  151. <!-- /results:bench -->
  152. <p align="center">
  153. <a href="https://ollaya.dev/results#speed"><img src="https://ollaya.dev/data/results/speed.svg" alt="Median latency of five-question requests for every model on each GPU and CPU we measured" width="760"></a>
  154. </p>
  155. ## Install
  156. - **Linux** (x86_64 or arm64, glibc ≥ 2.38, e.g. Ubuntu 24.04+):
  157. `curl -fsSL https://ollaya.dev/install.sh | sh`. When an NVIDIA GPU is present
  158. (driver R525+), the installer adds the CUDA runtime: CUDA 13 for R580+, CUDA 12 for older drivers.
  159. - **macOS** (Apple silicon): the same command.
  160. - **Windows** (x64): `irm https://ollaya.dev/install.ps1 | iex` in PowerShell. When an NVIDIA GPU is
  161. present (driver R527+), the installer adds the CUDA runtime, as on Linux.
  162. - **Desktop app** for macOS, Windows and Linux: start and stop the server, download models and run
  163. them in one window. On macOS it lives in the menu bar. Get it from
  164. [ollaya.dev/download](https://ollaya.dev/download).
  165. - **Docker:** `docker run -d --gpus=all -p 11435:11435 ghcr.io/ollaya-dev/ollaya:cuda`
  166. (`:cuda12` for host drivers older than R580), or `ghcr.io/ollaya-dev/ollaya` for CPU only.
  167. Configuration is through environment variables: `OLLAYA_HOST`, `OLLAYA_MODELS`,
  168. `OLLAYA_KEEP_ALIVE`, `OLLAYA_DEVICE`, `OLLAYA_API_KEY` and others, listed in
  169. [docs/api.md §15](docs/api.md).
  170. ## Repository
  171. | Path | What |
  172. |---|---|
  173. | `crates/ollaya` | The binary: CLI, daemon, runner |
  174. | `crates/ollaya-server` | HTTP API, scheduler (one runner process per model), model resolution |
  175. | `crates/ollaya-api` | API types and client; the contract is [docs/api.md](docs/api.md) |
  176. | `crates/ollaya-registry` | Model names, manifests, blob store, resumable pulls |
  177. | `crates/ollaya-decision` | Question schema, sequence layouts, calibration, answers |
  178. | `crates/ollaya-runner` | Inference engines (ONNX Runtime, and llama.cpp for GGUF models) |
  179. | `crates/ollaya-lang` | Script and language detection for routers |
  180. | `convert/` | Build-time Python: ONNX export, parity checks, packaging |
  181. | `site/` | The website and the static model registry host |
  182. ## Development
  183. ```sh
  184. cargo test --workspace
  185. cargo build --release -p ollaya --features cuda # CUDA build (x86-64 Linux and Windows)
  186. ```
  187. Tests come in three tiers, all with the same parity tolerances:
  188. - **K1, every change:** the tests, and smoke parity for the families the change touches. A parity
  189. example run with `--smoke` checks only the cases listed in
  190. [`crates/ollaya-runner/examples/smoke/cases.txt`](crates/ollaya-runner/examples/smoke/cases.txt),
  191. about 40 questions per family: every question type, refused requests, a long state and many options.
  192. ```sh
  193. cargo run --release -p ollaya-runner --example parity -- <model-dir> <goldens.jsonl> cpu --smoke
  194. ```
  195. - **K2, every release:** the full goldens on every GPU device a release ships, and on the CPU the full
  196. set for models of 4B and under and `--smoke` for larger ones. CPU goldens count only on the machine
  197. and build that made them.
  198. - **K3, major releases, a llama.cpp or ONNX Runtime update, a new device or a new model:** full CPU
  199. parity of the large models and the quality benchmarks. The site, this README and the reports are
  200. updated from those results.
  201. `convert/` rebuilds models. It exports them, checks parity against the PyTorch reference,
  202. generates golden fixtures, and packages the result into `registry/`. See the module docstrings.
  203. `cd convert && uv sync` installs it: with CUDA 13 torch on Linux and Windows, and with the CPU and
  204. MPS build from PyPI on Apple silicon Macs, where exports and parity run on the CPU.
  205. ## License
  206. Apache-2.0. Each model keeps its own license: `laya` (Convai Innovations), `decider` (Mapika),
  207. `kev` (Jared Palmer, on Qwen3.5 by the Qwen team), `decision` (the vLLM Semantic Router
  208. contributors, on Qwen3.5), `qwen3guard` (Qwen team), `gliclass` (Knowledgator), `von` (Victor Hugo
  209. Panisa), `winnow` (EldanRing, on Gemma 4 by Google DeepMind), `jevk5` (alibiserikbay, on Qwen3.5), `nimble`
  210. (Bespoke Labs, on Qwen3.5), `pplx-decider` (Perplexity, on Qwen3.5), `strands-decider` (Strands Agents, on Qwen3.5 and Gemma 4), `jeeves` (PostHog, on Qwen3.5), `clef` (Cloudflare, on Qwen3.5), `jeb` (Jason Brashear, on Qwen3.5 and Qwen3.8), `cygnet` (Gemma 4 by Google
  211. DeepMind; the Cygnet recipe is MIT), `snap` (logitlab, on MiniCPM5 by OpenBMB; snap's prompt is MIT), `decima` (A. M. Madani, on mmBERT-base by JHU CLSP and multilingual-e5-small by intfloat, both MIT), `largha` (Phocinae, on mmBERT-small by JHU CLSP, MIT) and `nli:modernbert-large` are Apache-2.0, and `nli:deberta-v3-large` (Moritz Laurer) is MIT. `arbiter`'s LoRA adapter and head (Codekins Pvt Ltd, Zyot Lab) are Apache-2.0; its base model, Gemma 3 4B IT by Google DeepMind, is under the Gemma Terms of Use. `nimble:9b-v3`'s adapter (Bespoke Labs) is CC BY-NC 4.0, which does not allow commercial use; its base, Qwen3.5-9B, is Apache-2.0. `d1` (Liquid AI) is under the LFM Open License v1.0: commercial use by an organization with US$10 million or more in annual revenue is not covered by it. llama.cpp, which Ollaya ships for
  212. GGUF models, is MIT.
  213. Ollaya is an independent project. It is not affiliated with or endorsed by Ollama or TypeSafe.

Ratings and comments

Sign in to rate this skill.

Sign in

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