Automated README summary
Automatically summarized from the README on 23 Oct 2026; not a run by jevbooks. Run instructions and author-reported outcomes still need verification.
What it is
token-saver is a Python CLI and library that picks the cheapest adequate Claude model and effort level for each prompt, so trivial requests don't get Opus-at-xhigh treatment. It targets developers integrating Claude who want cost-aware routing.
How it uses Jev
One Jev call asks four typed questions on the raw prompt: task_type (choice), difficulty (score, 0-4), deep_reasoning (noul), sensitive (noul). The answers become signals; a YAML policy ladder maps difficulty to a tier and adjusts it with confidence, reasoning and task-type rules, producing a model + effort decision used by ClaudeExecutor.
Primitives:choicescorenoul
Technique worth stealing
Ask several typed questions in one Jev call (they run in parallel) and use the returned confidence as a second routing axis.
Try it
pip install git+https://github.com/minBuild/token-saver.git, set TYPESAFE_API_KEY, then run token-saver route "...".
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-1.14.0- Project signals
jev-1.14.0- Category
jev-1.14.0- Pattern tags
jev-1.14.0
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 material0.95
- Shows a System One patternunclear0.52
- Describes handling uncertaintynot found in the excerpt0.04
- Reports measured numbersFound in material0.75
- Is a runnable toolFound in material0.86
- Top-10 learning shortlist signalunclear0.46
- Is a replica of the modelnot found in the excerpt0.02
- Solves a common problem (0–2)score on a 0–2 scale1.51 / 2
- About JevFound in material0.99
no-judgment band (0.40–0.60)
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 23 Oct 2026.