
Weco AI
β Text source-checked Nov 21, 2026Weco AI optimizes code against measurable metrics such as speed, accuracy, and cost.
Source-checked against the vendor's site Β· rebuilt from the vendor's own pagesProfile updated Link checked How we rate
In short
Weco AI optimizes code against user-defined metrics like speed and cost using a tree search structure. It integrates with Claude Code, Codex, and Cursor for iterative improvement.
Weco AI is a research and product lab building recursively self-improving AI. The Weco platform brings production-grade autoresearch to codebases, optimizing code against any metric defined by the user.
The optimization process uses a tree search structure where users can open any node to see its code, terminal output, and metric. Users can steer the run mid-process by trying a paper or their own idea, which becomes a new branch of the search. The system is model-neutral and supports GPT, Claude, Gemini, Kimi, GLM, and MiniMax with a bring-your-own-key option.
Users can set up optimization by typing /weco in Claude Code, Codex, or Cursor, where the coding agent sends the task to Weco, which iteratively writes the next code version and runs the evaluation.
According to Weco AI: Weco AI is a research and product lab building recursively self-improving AI. Its pages list this strength: The search process is observable, Allowing users to see code, terminal output, and metrics for any node.
β The AI Tool Bible editorial team
Pros
- β The search process is observable, allowing users to see code, terminal output, and metrics for any node.
- β The system is steerable mid-run, allowing users to add new branches based on papers or their own ideas.
- β The platform is model-neutral, supporting multiple LLM providers with a bring-your-own-key option.
- β Weco integrates directly with popular coding agents like Claude Code, Codex, and Cursor.
Facts
- Platforms
- cli, api (not source-checked)
- Pricing
- freemium (not source-checked)
- API available
- Yes (not source-checked)
- Free trial
- Yes (not source-checked)
- Model used
- Multi-model (LLM + AIDE tree search) (not source-checked)
- Hacker News
- 5 stories Β· top 10 points Β· last Jul 15, 2026 Β· checked Oct 23, 2026
- Earliest Wayback capture
- 2025-03-19 (earliest archived capture of the site, not a launch date)
Where this tool sits
Use cases
Frequently asked
- How much does Weco AI cost?
- Weco AI operates on a freemium pricing model. Specific pricing details are not documented in the provided information, so users should check the official platform for current tier costs and limits.
- Which coding agents does Weco AI integrate with?
- Users can set up optimization by typing /weco in Claude Code, Codex, or Cursor. The coding agent sends the task to Weco, which then iteratively writes the next code version and runs the evaluation.
- What AI models does Weco AI support?
- The system is model-neutral and supports GPT, Claude, Gemini, Kimi, GLM, and MiniMax. It also offers a bring-your-own-key option, allowing users to utilize their preferred model providers for the optimization process.
- How does the Weco AI optimization process work?
- It uses a tree search structure where users can inspect code, terminal output, and metrics at any node. Users can steer the run mid-process by introducing a new idea or paper, which becomes a new branch of the search.
- What types of code optimization can Weco AI handle?
- Weco AI optimizes code against any metric defined by the user, such as speed, accuracy, and cost. Use cases include fraud-detection, prompt-engineering, inference-optimization, hyperparameter-optimization, and general code-optimization.