designcomputer/systemone-client
User repository on GitHub
Python client for the Ollama System One decision endpoint
- Forks
- 0
- Last pushed
- Updated Oct 2026
Repo stats from the GitHub API, cached Oct 2026.
Jev directory
CommunityJev Repos & SDKs
Systemone Client
A one-file Python client for the Ollama System One decision endpoint that adds image judging, fallback of uncertain answers to a stronger model and pre-flight size checks on top of what the official ollama package offers, with requests as its only dependency.
- Stars
- 0
- Language
- Python
- Reading time
- 2 min
- Published by
- Community
- Author
- designcomputer
- Last verified
- Oct 2026
Used inJev for Software & DevTools
Highlights
- Python client for Ollama's System One decision endpoint: classify, answer yes/no or score against a rubric in one pass.
- Typical server latency is about 70 to 500 ms depending on model and state size, with no generation or streaming.
- Adds image judging via images=[...] for Clef and Clef Flash, which the official ollama 0.6.3 package lacks.
- systemone_fallback() re-runs uncertain answers on a stronger model; oversized bodies fail locally before any request.
- One file and the requests library; no async client, and it can wrap ollama.systemone instead of replacing it.
Watch out
README and GitHub metadata reviewed 2026-10-17; the captured excerpt truncates in the fallback section, so installation, the typed answer shapes, the fallback trigger and any benchmarks beyond the latency range are not visible; no license is listed.
Questions about this entry
Short answers drawn from the linked project or page.
A Python client for the Ollama System One decision endpoint, POST /v1/systemone. System One models classify text, answer yes/no questions or score text against a rubric in a single forward pass, returning one JSON response with no generation or streaming.
The official package has supported System One since 0.6.3, but this client adds image judging for Clef and Clef Flash via images=[...], fallback escalation of uncertain answers through systemone_fallback(), and pre-flight checks on empty or oversized payloads.
The README says empty states and bodies over the limits, naming 64 KiB and 32 MiB, raise ValueError before any round trip, so bad input is rejected locally instead of costing a request.
Typical latency on the server is about 70 to 500 ms depending on the model and the size of the state; the README frames it as one forward pass rather than a generation loop.
Yes, systemone_fallback() re-runs uncertain answers on a stronger model, and the README notes you can pass the official ollama.systemone as the client rather than choosing one implementation or the other.
requests and a single file, against httpx and pydantic for the official package. It returns typed answers including score legend, probabilities and confidence; there is no async client, which the comparison table lists as the official package's advantage.
The captured excerpt cuts off in the middle of the fallback section, so installation steps, the fallback trigger, the full API surface and any benchmarks beyond the 70 to 500 ms range are unavailable; no license is listed.
More Jev resources like this
A System One layout optimizer for decision-model workloads: it reorders rows and fields so a prefix-caching backend reuses input computation, then returns per-record decisions in original order. Its bundled backend serves AutoTrust JEV-9B through vLLM.
Jev Repos & SDKs#community#github#python
Community
0814wdwd/solo_jevPython
0814wdwd/solo_jevPythonA Python library that turns Jev-style decision models into ordinary function calls: check returns a bool, classify picks one option, label picks every option that applies, and score rates on a scale, with probabilities behind every answer.
Jev Repos & SDKs#community#github#python
Community
jlowin/vibecheckPython
jlowin/vibecheckPythonHTTP server for the Mica v0.5 4B Preview decision model: GGUF weights run through llama.cpp and answer noul, choice and score questions with calibrated probabilities over the TypeSafe /v1/systemone format.
Jev Repos & SDKs#community#github#python
Community
akivet/MicaPython
akivet/MicaPythonServing code for the Jeff 1.0 family of Jev-style decision models: a state, a typed question and its options go in, and one forward pass returns a probability for each option. Four open models from 4B to 31B parameters are served the same way.
Jev Repos & SDKs#community#github#python
Community
j-geuter/jeff-servePython
j-geuter/jeff-servePythonA frozen decision server that serves Gemma 4 31B — quantized to FP8 at startup — in TypeSafe's /v1/systemone format, reading the next-token distribution instead of generating tokens. It needs one FP8 NVIDIA GPU with 80 GB and has no authentication.
Jev Repos & SDKs#community#github#python
Community
krishna-gogineni-765/deck31bPython
krishna-gogineni-765/deck31bPythonAn open-source, locally executed decision engine in the spirit of Jev (TypeSafe AI): state becomes features, and Noul, Choice and Score heads return typed, calibrated probabilities without generating text. It ships a cheap-to-strong cascade and a full train-calibrate-evaluate-deploy cycle.
Jev Repos & SDKs#community#github#python
Community
slavabars/decision-cascadePython
slavabars/decision-cascadePythonFrom the Jev guides
Original write-ups on what this entry is for.
From the Jev AI model community
Posts from builders shipping with Jev right now.