Jev Bisect

Jev Bisect is a Python library that finds numeric answers through repeated higher, lower, or exactly decisions using Jev.

Jev Bisect is a Python library designed to find numeric answers through repeated higher, lower, or exactly decisions using the official TypeSafe Python SDK. Users supply a question and bounds, and the library calculates guesses until Jev selects exactly. It features configurable precision and turn limits, and returns the final answer, turns, state, and history. The tool runs on Python 3.10 or newer and requires finite numbers for bounds. It also supports manual step advancement without an API call and allows users to reuse a client or pin a specific model.

Jev Bisect at a glance

Best for: Software engineer
Runs on: CLI
Pricing: Free
Released: October 2026
Category: AI coding tool
Website: github.com/mkly/jev-bisect

Key features

Finds numeric answers using higher, lower, or exactly decisions
Uses exact rational arithmetic and integer candidate indexes
Supports configurable precision and maximum turn limits
Returns answer, turns, final state, and history upon completion
Allows advancing search steps manually without an API call
Supports custom starting guesses and client reuse

Use it for

Finding precise numeric values within defined bounds
Automating interactive guessing games or searches
Running deterministic tests without an API key

Alternatives to Jev Bisect

jevgrep — Reduces coding agent costs by providing targeted, relevant repository context.
JevStation — Get typed decisions with probabilities instead of prose you have to parse.
Jevcrawl — Finds answer pages more often and cheaper than letting Claude crawl.
Jevia — Routes coding tasks to the right model tier using verified outcomes.
ej — Runs locally on CPU from a single 11 MB file without APIs or GPUs.
zvec — It is a pure local, serverless vector database providing fast hybrid search.
Jev Doc Search — It uses hierarchical tree search without vector databases or embeddings.
Fast Jev Compaction — It keeps tool calls and results verbatim using Jev decisions instead of summaries.
pstack — Improves AI coding agent outcomes with rigorous workflows.
Chelis — Catches shape and precision errors before code runs.
Jev State — Test conversational decisions and export runnable TypeScript tests
Simple Checker V1 — Verifies AI code with color-coded flowcharts and runtime evidence.

All Jev Bisect alternatives