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.
Across 348 long-horizon benchmark sessions, Tura used up to 83.1% fewer turns on the rewrite benchmark and improved the DeepSWE pass rate by up to 16.7 percentage points compared with Codex CLI.
| Date | Stars |
|---|---|
| 2026-07-31 | 493 |
| 2026-08-06 | 493 |
Today
— stars today
This week
— stars this week
This month
— stars this month
Momentum
35.0
growth rate 0.00%/day
<p align="center">
<a href="https://turaai.net/">
<img src="assets/tura/icon.svg" alt="Tura icon" width="96">
</a>
</p>
<p align="center">
<a href="https://turaai.net/"><img alt="Website" title="Tura official website" src="https://img.shields.io/badge/Website-turaai.net-40e0d0?style=flat-square&labelColor=555555"></a>
<a href="https://turaai.net/benchmark"><img alt="Benchmark: 8,243 turns" title="Tura benchmark: 8,243 agent turns" src="https://img.shields.io/badge/Benchmark-8%2C243_turns-9b59b6?style=flat-square&labelColor=555555"></a>
<a href="https://www.npmjs.com/package/tura-ai"><img alt="npm package" title="Tura npm package" src="https://img.shields.io/npm/v/tura-ai?style=flat-square&logo=npm&label=npm&labelColor=555555&color=cb3837"></a>
</p>
<p align="center"><strong>English</strong> | <a href="README.zh-CN.md">简体中文</a></p>
<h1 align="center">Tura: 16.7% better performance, 77.5% fewer rounds.</h1>
Tura is a local, open-source coding agent for developers who are tired of vague skill claims, token-saving extensions with no evidence, and agents that change a repository before understanding it.
Across 20 DeepSWE v1.1 tasks, each run three times per agent, Tura creates a substantial token-budget advantage by reducing repeated context and model round trips. You can spend that advantage in two ways. Direct turns most of it into lower cost: 77.5% fewer aggregate tokens than Codex CLI, with a comparable verifier success rate of 65.0% versus 63.3%. Balanced puts more of the saved budget back into reasoning, investigation, and verification. It reached an 80.0% success rate—16.7 percentage points higher than Codex CLI—while still using 31.1% fewer tokens.[^debug-figure][^debug-manifests]
### Benchmark
Long-horizon task [benchmarks](https://turaai.net/benchmark) are one way to look past a polished isolated prompt and see how an agent handles real work. The published comparison uses harness-based development tasks with archived prompts, per-round tool calls, token usage, patches, and verifier results.
> The published artifacts compare the named Tura Balanced, Tura Direct, and Codex CLI configurations on 20 DeepSWE tasks, 5 rewrite tasks, and 2 separately reviewed design tasks. [^debug-figure]. [current test-set record](https://github.com/Tura-AI/benchmark/blob/main/doc/current-test-set-record.md). [^test-set-record]
The published results do not establish equivalent quality or performance for
every configured provider. Broader Anthropic/Claude, Google/Gemini,
OpenAI-compatible, local-provider, UI-latency, runtime/session parsing, and
cross-OS measurements remain part of the documented
[roadmap](ROADMAP.md) and [known evidence gaps](docs/KNOWN_ISSUES.md).
<details>
<summary><strong>FULL BENCHMARK REPORT</strong></summary>
<p align="center">
<img src="assets/data/benchmark-agent-comparison.svg" alt="DeepSWE Debug and Rewrite Repo benchmark comparison" width="800">
</p>
<p align="center"><em>Harness success and aggregate token usage across 25 high-difficulty tasks, 6 agent-and-model configurations, and 270 sessions. Source and calculation notes are linked below.</em></p>
</details>
### Screenshots
<p align="center">
<img src="assets/screenshot/gui-ci-quality-demo.svg" alt="Tura GUI" width="800">
</p>
<p align="center"><em>GUI page with multi-session concurrent work and HTML rich text support.</em></p>
<p align="center">
<img src="assets/screenshot/tui-ci-quality-demo.svg" alt="Tura TUI" width="800">
</p>
<p align="center"><em>TUI page with multi-session concurrent work and HTML rich text support.</em></p>
The results below come from published benchmark artifacts, not an uncited aggregate. Three systems do most of the work:
## Macro CLI Command Run
Most coding agents still depend on repetitive tool-calling loops: inspect, wait, patch, wait, build, wait, test, wait.
_**Tool-calling coding agent:**_
```bash
# Turn 1 — inspect environment
rg -n "TODO|command_run|handler" craExcerpt of 21,069 characters
Read on GitHubWould you bet a product on this? Bounded 0–100 and slow moving.
matched fp:ed23f525b1e0b39f, llm:Repository topics: agent, agentic-ai, coding-agent, context-engineering, developer-tools-ai-agent, llm, terminal-based, token-optimization; description mentions benchmark sessions and improvements vs Codex CLI
matched fp:ed23f525b1e0b39f, llm:Repository topics: agent, agentic-ai, coding-agent, context-engineering, developer-tools-ai-agent, llm, terminal-based, token-optimization; description mentions benchmark sessions and improvements vs Codex CLI
matched fp:ed23f525b1e0b39f, llm:Repository topics: agent, agentic-ai, coding-agent, context-engineering, developer-tools-ai-agent, llm, terminal-based, token-optimization; description mentions benchmark sessions and improvements vs Codex CLI