Ollaya

ollaya-dev1.4kstars

1# Ollaya

README.mdMCP server

Transport: stdio
Status: Local

Power your agent

Add this server

Install from the publisher. Not a PlugMCP endpoint.

Run on PlugMCP

Host a copy with our keys and logs.

No packaged release in the registry yet. Follow the repository README.

RepositoryPublisher installStep-by-step: add this in Claude

Remote URL or a command on your machine? How the two installs differ

Run on PlugMCP

Host a copy with our keys and logs.

A new server you create. It is not this listing's official endpoint. Use the publisher install when you want their URL and their credentials.

Launch

Server facts

Local

Transport: stdio
License: Apache-2.0
Last checked: 2026-10-27

Stars: 1.4k
Tools: —
Uptime: —

Overview

About Ollaya

Jev-compatible API server for open decision models. Ollama for decision models: pull and serve 25 model families (Laya, pplx-decider, Clef, Winnow, coba...) locally or self-hosted behind /v1/systemone. The TypeSafe SDK works unchanged: set its base URL and pick an open model. CLI, desktop app, Docker, MCP server, public bench.

Tools

No tool list yet. Remote servers get one after the first health check.

FAQ

Questions

How do I install Ollaya?

Copy the stdio config for Cursor, Claude Desktop, Claude Code, or VS Code on this page. It uses the publisher's install, not a PlugMCP proxy.

Does Ollaya need an API key?

No required environment variable is listed on this record.

What transport does Ollaya use?

Ollaya runs locally over stdio.

When was Ollaya last checked?

Last checked 2026-10-27. Jev-compatible API server for open decision models. Ollama for decision models: pull and serve 25 model families (Laya, pplx-decider, Clef, Winnow, coba...) locally or self-hosted behind /v1/systemone. The TypeSafe SDK works unchanged: set its base URL and pick an open model. CLI, desktop app, Docker, MCP server, public bench.

From the publisher

Documentation

curl -fsSL https://ollaya.dev/install.sh | sh
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."

intent            refund                                ███████████████░ 0.91
is_urgent         yes                                   ███████████████░ 0.92
frustration       2.89 / 3  very angry or using stron…  ██████████████░░ 0.86
refund_requested  yes                                   ████████████████ 0.99
churn_risk        yes                                   ████████████████ 0.99

Or ask your own questions: ollaya run kev --ask "Is a refund requested?" --ask "Which team?" --choice billing --choice sales < ticket.txt.

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.

A self-hosted, Jev-compatible API

What is a Jev-compatible API? An HTTP API that takes the same requests as TypeSafe's hosted Jev (a state plus typed questions at POST /v1/systemone) and returns the same answers, so TypeSafe's SDK and any client written for Jev work against it unchanged. Ollaya's /v1/systemone, /v1/decisions and /v1/models are wire-identical to TypeSafe's (as typesafe-sdk 0.7.1 defines them), served from your own machine with open models.

Migrating from Jev. Point the SDK at Ollaya and pick an open model:

export TYPESAFE_BASE_URL=http://localhost:11435
export TYPESAFE_API_KEY=local              # any non-empty value
export TYPESAFE_DEFAULT_MODEL=winnow:e4b   # instead of "jev-latest"

Details, including proxies and warm-up: TypeSafe compatibility.

	Ollaya	TypeSafe Jev	Ollama
Runs on	Your machine or server	TypeSafe's cloud	Your machine or server
Jev API	`/v1/systemone`, `/v1/decisions`, `/v1/models`	The original	`/v1/systemone`
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)
Probabilities	Each author's calibration	TypeSafe's	Raw softmax
Your data leaves the machine	No	Yes	No
License	Apache-2.0	Proprietary, hosted	MIT

The FAQ compares Ollaya and Ollama in more detail, measured on the same GPU.

Features

One binary. ollaya serve runs the daemon; ollaya run, pull, list, ps, show, rm, cp, stop and create work the way they do in Ollama. If the daemon isn't running, the CLI starts it.
TypeSafe-compatible. POST /v1/systemone, /v1/decisions and GET /v1/models are wire-identical to TypeSafe. The official SDK works unchanged when you set TYPESAFE_BASE_URL=http://localhost:11435.
Native API. /api/decide adds routing information and timings. /api/pull streams NDJSON progress, and there are /api/tags, /api/show, /api/ps and more. See docs/api.md.
Weights come from their authors. Ollaya publishes only small ONNX graphs, about 3 MB each. These graphs read the original weight files (usually model.safetensors) from the author's Hugging Face repository, pinned to a commit and verified by sha256. Models whose authors publish GGUF files (winnow, jevk5, jeb) run that file itself on llama.cpp. Ollaya never re-hosts weights.
For agents. ollaya mcp serves the models to Claude Code, Claude Desktop, Cursor and other MCP clients (claude mcp add ollaya -- ollaya mcp, or one click in Cursor), and the ollaya-decisions skill teaches agents when and how to use them (npx skills add ollaya-dev/ollaya --skill ollaya-decisions).
Routers. laya detects the script and language of each request, then answers with laya:en or laya:multilingual.
Modelfiles. You can bake a question set into your own model:
```
FROM laya
QUESTIONS ./triage.json
PARAMETER precision fp32
```
Then run ollaya create triage -f Modelfile and ollaya run triage "…".
Fast and exact.
- Hardware: ONNX Runtime on CPU, and CUDA on NVIDIA GPUs. GGUF models run on llama.cpp: 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.
- Precision: ONNX uses fp16 on GPU and fp32 on CPU when available; GGUF keeps its model quantization on either device.
- Accuracy: fp32 exports give the same decision as the PyTorch reference on 100% of 2,383 questions per checkpoint.

Models

Model	What it is
`winnow:e4b`	Recommended. EldanRing's Winnow-E4B, a Gemma 4 fine-tune run from the author's Q8_0 GGUF on llama.cpp
`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
`laya`	Router: picks `laya:en` or `laya:multilingual` by language
`laya:en`	English decision model (ModernBERT-large, 421M)
`laya:multilingual`	100+ languages (mmBERT-base, 322M)
`laya:typed-decisions`	Fine-tuned on the typed-decisions workflows
`decider`, `decider:4b`, `decider:0.8b`	Mapika's Qwen3.5 decoders, 2B (the default), 4B and 0.8B
`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
`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
`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
`qwen3guard`	Qwen3Guard-Gen-0.6B safety guard with built-in questions: safe, controversial or unsafe, and the category
`nli`, `nli:modernbert-large`	Moritz Laurer's zero-shot NLI classifiers (DeBERTa-v3-large, ModernBERT-large)
`gliclass`	Knowledgator's instruction-following zero-shot classifier (DeBERTa-v3-large)
`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
`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
`winnow`	EldanRing's Winnow-12B, the larger sibling of `winnow:e4b`
`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
`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
`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
`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
`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
`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
`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
`jeeves`	PostHog's Jeeves-9B without its reasoning chain: Qwen3.5-9B (LoRA merged) and a pointer head
`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
`cygnet`	blockbrain-ai's Cygnet: frozen Gemma 4 12B IT (Q8_0 GGUF) with a letter-readout prompt and temperature 3.4
`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
`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
`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
`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
`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
`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

Browse them at ollaya.dev/search; every tag's accuracy, speed and memory is on Ollaya Bench. Decoder tags ending in -fp16 (kev:4b-fp16) compute in fp16 on NVIDIA GPUs: faster, with at least 99.5 % of their decisions the same as the base tag's. Laya's -fp32 and -fp16 tags pin the precision. The derived files of every model are also published at huggingface.co/ollaya-dev.

Ollama 0.40 also serves decision models, nine of them, through the same TypeSafe wire format (measured side by side). The FAQ compares the two projects.

Results

Every model in the library is measured on our own machines and published, with the raw data, on Ollaya Bench: one released runtime, each author's own weights, the same 5,880 questions for every model (Bespoke Labs' 13 public subsets and typed-decisions), and speed and memory on a GPU and a CPU.

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:

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.
Calibration. The lowest calibration errors on the public benchmark are 0.016 (pplx-decider:flash), 0.018 (jeeves:9b) and 0.020 (jeb:9b).
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.
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.

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.

Install

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.
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.

Repository

Path	What
`crates/ollaya`	The binary: CLI, daemon, runner
`crates/ollaya-server`	HTTP API, scheduler (one runner process per model), model resolution
`crates/ollaya-api`	API types and client; the contract is docs/api.md
`crates/ollaya-registry`	Model names, manifests, blob store, resumable pulls
`crates/ollaya-decision`	Question schema, sequence layouts, calibration, answers
`crates/ollaya-runner`	Inference engines (ONNX Runtime, and llama.cpp for GGUF models)
`crates/ollaya-lang`	Script and language detection for routers
`convert/`	Build-time Python: ONNX export, parity checks, packaging
`site/`	The website and the static model registry host

Development

cargo test --workspace
cargo build --release -p ollaya --features cuda   # CUDA build (x86-64 Linux and Windows)

Tests come in three tiers, all with the same parity tolerances:

K1, every change: the tests, and smoke parity for the families the change touches. A parity example run with --smoke checks only the cases listed in crates/ollaya-runner/examples/smoke/cases.txt, about 40 questions per family: every question type, refused requests, a long state and many options.
```
cargo run --release -p ollaya-runner --example parity -- <model-dir> <goldens.jsonl> cpu --smoke
```
K2, every release: the full goldens on every GPU device a release ships, and on the CPU the full set for models of 4B and under and --smoke for larger ones. CPU goldens count only on the machine and build that made them.
K3, major releases, a llama.cpp or ONNX Runtime update, a new device or a new model: full CPU parity of the large models and the quality benchmarks. The site, this README and the reports are updated from those results.

convert/ rebuilds models. It exports them, checks parity against the PyTorch reference, generates golden fixtures, and packages the result into registry/. See the module docstrings. cd convert && uv sync installs it: with CUDA 13 torch on Linux and Windows, and with the CPU and MPS build from PyPI on Apple silicon Macs, where exports and parity run on the CPU.

License

Apache-2.0. Each model keeps its own license: laya (Convai Innovations), decider (Mapika), kev (Jared Palmer, on Qwen3.5 by the Qwen team), decision (the vLLM Semantic Router contributors, on Qwen3.5), qwen3guard (Qwen team), gliclass (Knowledgator), von (Victor Hugo Panisa), winnow (EldanRing, on Gemma 4 by Google DeepMind), jevk5 (alibiserikbay, on Qwen3.5), nimble (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 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 GGUF models, is MIT.

Ollaya is an independent project. It is not affiliated with or endorsed by Ollama or TypeSafe.

intent refund ███████████████░ 0.91 is_urgent yes ███████████████░ 0.92 frustration 2.89 / 3 very angry or using stron… ██████████████░░ 0.86 refund_requested yes ████████████████ 0.99 churn_risk yes ████████████████ 0.99

Ollaya

TypeSafe Jev

Ollama

Runs on

Your machine or server

TypeSafe's cloud

Your machine or server

Jev API

/v1/systemone, /v1/decisions, /v1/models

The original

/v1/systemone

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)

Probabilities

Each author's calibration

TypeSafe's

Raw softmax

Your data leaves the machine

Yes

License

Apache-2.0

Proprietary, hosted

MIT

Model

What it is

winnow:e4b

Recommended. EldanRing's Winnow-E4B, a Gemma 4 fine-tune run from the author's Q8_0 GGUF on llama.cpp

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

laya

Router: picks laya:en or laya:multilingual by language

laya:en

English decision model (ModernBERT-large, 421M)

laya:multilingual

100+ languages (mmBERT-base, 322M)

laya:typed-decisions

Fine-tuned on the typed-decisions workflows

decider, decider:4b, decider:0.8b

Mapika's Qwen3.5 decoders, 2B (the default), 4B and 0.8B

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

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

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

qwen3guard

Qwen3Guard-Gen-0.6B safety guard with built-in questions: safe, controversial or unsafe, and the category

nli, nli:modernbert-large

Moritz Laurer's zero-shot NLI classifiers (DeBERTa-v3-large, ModernBERT-large)

gliclass

Knowledgator's instruction-following zero-shot classifier (DeBERTa-v3-large)

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

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

winnow

EldanRing's Winnow-12B, the larger sibling of winnow:e4b

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

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

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

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

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

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

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

jeeves

PostHog's Jeeves-9B without its reasoning chain: Qwen3.5-9B (LoRA merged) and a pointer head

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

cygnet

blockbrain-ai's Cygnet: frozen Gemma 4 12B IT (Q8_0 GGUF) with a letter-readout prompt and temperature 3.4

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

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

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

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

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

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

Path

What

crates/ollaya

The binary: CLI, daemon, runner

crates/ollaya-server

HTTP API, scheduler (one runner process per model), model resolution

crates/ollaya-api

API types and client; the contract is docs/api.md

crates/ollaya-registry

Model names, manifests, blob store, resumable pulls

crates/ollaya-decision

Question schema, sequence layouts, calibration, answers

crates/ollaya-runner

Inference engines (ONNX Runtime, and llama.cpp for GGUF models)

crates/ollaya-lang

Script and language detection for routers

convert/

Build-time Python: ONNX export, parity checks, packaging

site/

The website and the static model registry host