Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
The agent framework where every run is durable, replayable, and resumable by default.
| Date | Stars |
|---|---|
| 2026-07-31 | 1361 |
| 2026-08-04 | 1362 |
| 2026-08-06 | 1362 |
Today
— stars today
This week
— stars this week
This month
— stars this month
Momentum
35.0
growth rate 0.00%/day
<p align="center"> <img src=".github/chidori-banner.svg" alt="Chidori — checkpoint · replay · resume: durable TypeScript agents on a Rust core" width="800" /> </p> <h1 align="center">Chidori</h1> <p align="center"> <b>The agent framework where every run is durable, replayable, and resumable by default.</b> </p> <p align="center"> Write agents as plain async TypeScript. Every side effect — every LLM call, tool call, and HTTP request — flows through the runtime as a recorded <b>host call</b>. So any run can be checkpointed to disk, <b>replayed for byte-identical output with zero LLM calls</b>, and resumed from any pause — even in a new process after a crash. One Rust binary, an embedded pure-Rust JavaScript engine, and TypeScript + Python SDKs. No Node, no DSL, no native bindings. </p> <p align="center"> <a href="https://github.com/ThousandBirdsInc/chidori/commits"><img alt="GitHub Last Commit" src="https://img.shields.io/github/last-commit/ThousandBirdsInc/chidori" /></a> <a href="https://crates.io/crates/chidori"><img alt="crates.io version" src="https://img.shields.io/crates/v/chidori" /></a> <a href="https://pypi.org/project/chidori/"><img alt="PyPI version" src="https://img.shields.io/pypi/v/chidori" /></a> <a href="https://www.npmjs.com/package/@1kbirds/chidori"><img alt="npm version" src="https://img.shields.io/npm/v/%401kbirds%2Fchidori" /></a> <a href="https://github.com/ThousandBirdsInc/chidori/blob/main/LICENSE"><img alt="License Apache-2.0" src="https://img.shields.io/badge/License-Apache_2.0-blue.svg" /></a> </p> <p align="center"> <a href="#-why-chidori"><b>💡 Why Chidori</b></a> · <a href="#️-quick-start"><b>⚡️ Quick Start</b></a> · <a href="#-what-you-can-build"><b>🧰 What You Can Build</b></a> · <a href="#️-how-chidori-compares"><b>⚖️ Compare</b></a> · <a href="#-documentation"><b>📚 Docs</b></a> · <a href="https://discord.gg/CJwKsPSgew"><b>💬 Discord</b></a> </p> ## 💡 Why Chidori **Agents are non-deterministic, expensive, and long-running.** That combination is what makes them miserable to build: - 🐛 A bug surfaces three runs deep — and you can't reproduce it. - 💸 Every debugging cycle re-bills the same tokens. - 💥 A crash halfway through a multi-step run loses everything. - ⏳ "Wait for a human to approve" means keeping a process alive for hours. Most frameworks layer orchestration *on top of* this chaos. **Chidori removes it at the source.** The trick is a single boundary. Every side effect an agent performs — every LLM call, tool call, and HTTP request — flows through the runtime as a recorded **host call**. Agents never touch the world directly, so the runtime sees (and records) *everything*: <p align="center"> <img src=".github/host-call-loop.svg" alt="Animation: a TypeScript agent calls host functions on the Chidori runtime; the runtime performs each side effect against the world (LLMs, HTTP, tools) and records it in the call log" width="860" /> </p> Once the runtime sees every side effect, it can log it, cache it, replay it, pause on it, and resume from it. That one mechanism is what turns each of the four problems above into a feature: - 🔁 **Replay any run with zero LLM calls.** The call log is a deterministic record. Re-run the exact same code against it — for tests, for debugging, for recovery — and every prompt, tool, and HTTP call returns its recorded result instantly. No tokens spent, identical output. - 💾 **Survive crashes and restarts.** Runs are checkpointed at every host safepoint. Kill the process mid-run and resume exactly where it left off — in a brand-new process — by replaying the call log to the pause point and continuing live. - 🧑⚖️ **Pause for humans without holding a process open.** `chidori.input()` and named [signals](./docs/signals.md) suspend the run to disk. A human (or another agent) answers minutes or days later and the run picks up exactly where it stopped. - 🧪 **Check in a checkpoint as a test.** Commit a recor
Excerpt of 24,090 characters
Read on GitHub
314
6
1
1
Would you bet a product on this? Bounded 0–100 and slow moving.
matched fp:0ecea2e4a901780f, topic:agents, topic:agent-framework, desc:agent framework