# ThousandBirdsInc/chidori

The agent framework where every run is durable, replayable, and resumable by default.

Repository: https://github.com/ThousandBirdsInc/chidori
Canonical: https://ross.abutalabs.com/products/chidori
Homepage: https://thousandbirdsinc.github.io/chidori/
Language: Rust
License: Apache-2.0
License Family: permissive
Topics: agents, ai, agent-framework, checkpointing, deterministic, durable-execution, javascript-engine, llm, replay, resumable, rust, typescript
Last push: 2026-08-26T10:57:26+00:00

## Health v2 (maintenance only)
Score: 94/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 95, longevity 81
- inputs: {"age_days": 1141, "days_push": 7, "days_rel": 35, "gap_med": 17, "n_releases_24m": 4}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1363, forks 59 (observed 2026-08-28T04:04:30.376108+00:00)

## What it is
Chidori is an agent framework that runs async TypeScript agents on an embedded pure-Rust JavaScript engine, recording every side effect (LLM calls, tool calls, HTTP requests) as host calls. This makes every run durable, checkpointable to disk, byte-identically replayable with zero LLM calls, and resumable across process crashes.

## Use cases
- build AI agents that survive crashes and resume mid-run
- replay an agent run deterministically without spending tokens on LLM calls
- pause an agent run for human approval and resume days later
- use recorded agent runs as free, fast integration tests
- debug a multi-step agent run that failed deep in execution
- run TypeScript agents without Node or V8 using a single Rust binary

## When to choose
- you need deterministic replay and checkpointing for long-running or expensive LLM agent workflows
- you want human-in-the-loop pauses that persist to disk without keeping a process alive
- you want to commit recorded agent runs as regression tests
- you prefer a dependency-free Rust binary over Node-based agent frameworks

## When to avoid
- you need a mature ecosystem with extensive integrations and a large community
- your agents must perform side effects outside the recorded host-call boundary
- you primarily work in languages other than TypeScript or Python

## Facets
- artifact type: framework
- maturity: active
- function: agent-framework, llm-inference, sdk, testing, caching
- domain: large-language-models, developer-tools
- platform: rust, python, cross-platform, cli
- tags: durable-execution, checkpointing, replay, resumable, deterministic, javascript-engine, host-calls, prompt-caching, ai-agents, automation, typescript

## Member repositories
- ThousandBirdsInc/chidori (main) score 94

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:30.376108+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T04:41:24.020236+00:00, confidence not recorded.
  - readme: https://github.com/ThousandBirdsInc/chidori (fetched 2026-08-28T04:04:30.376108+00:00, sha 2290c9ca9454)
  - homepage: https://thousandbirdsinc.github.io/chidori/ (fetched 2026-08-29T11:58:51.353046+00:00, sha 0b5642e9d764)
  - registry_crates: https://crates.io/api/v1/crates/chidori (fetched 2026-08-29T11:58:51.362640+00:00, sha 22aab6dd7b48)
- Data as of 2026-08-30T08:39:29.467469+00:00.