jevbooks

← All projects

tachyone

munod/tachyone · Homepage

Local-first multilingual System One decision engine speaking the Jev protocol.

Agent Decisions0.41Runner-up: Infra / SDKs / Integrations
Automated README summary

Automatically summarized from the README on 19 Oct 2026; not a run by jevbooks. Run instructions and author-reported outcomes still need verification.

What it is

Tachyone is a self-hosted decision engine for developers who need fast, calibrated atomic judgments instead of generated text. It answers choice, score and noul questions over a given state, and is a drop-in replacement for the Jev /v1/systemone wire contract.

How it uses Jev

Tachyone implements the Jev /v1/systemone protocol: a client sends a state plus typed questions (choice, score, noul) and receives typed answers with probabilities and calibrated confidence. The local encoder backend runs one non-autoregressive forward pass (ModernBERT/mmBERT plus LoRA adapters) to produce those decisions.

Primitives:choicescorenoul

Technique worth stealing

Non-autoregressive single forward pass with strictly proper scoring (RLCD) and per-(primitive, language) temperature calibration.

Try it

uv sync; uv run tachyone --predict --preset triage --backend fake "refund please"; or run uv run tachyone-serve.

View on GitHub

Start using Jev: API key and your first request →

Model judgments from README excerpts

Project signals use repository metadata, an existing summary and the first 700 README characters, not the full project. “Not found” means not detected in that material; it does not establish absence. These are the original judgments with recorded models and dates, not source audits or run results.

Models & evaluation dates

Inclusion gate
jev-1.13.0
Project signals
jev-1.13.0
Category
jev-1.13.0
Pattern tags
jev-1.13.0

Evidence

These signals use repository metadata, an existing summary and at most 700 cleaned README characters. ≥ 0.60 means reported in that excerpt; ≤ 0.40 means not found there. The middle band is inconclusive. None is a source-code audit or a test result. The runnable-tool question excludes research experiments; the recommendation question asks for a learning shortlist, not a quality verdict.

  • Jev is core to the projectFound in material0.92
  • Shows a System One patternnot found in the excerpt0.28
  • Describes handling uncertaintynot found in the excerpt0.06
  • Reports measured numbersnot found in the excerpt0.06
  • Is a runnable toolFound in material0.93
  • Top-10 learning shortlist signalunclear0.40
  • Is a replica of the modelFound in material0.86
  • Solves a common problem (0–2)score on a 0–2 scale1.15 / 2
  • About JevFound in material0.97
Signals
Jev is core to the project0.92Shows a System One pattern0.28Describes handling uncertainty0.06Reports measured numbers0.06Is a runnable tool0.93Top-10 learning shortlist signal0.40Is a replica of the model0.8600.51Solves a common problem (0–2)1.15 / 2012

no-judgment band (0.40–0.60)

Design patterns
Questionnaire0.68Funnel0.62Check-up0.48Grey zone0.31Traffic light0.2000.51

no-judgment band (0.40–0.60)

The Jev model and date for each judgment stage are listed above. Card written by DeepSeek V4.1 Flash from the README on 19 Oct 2026.