jcode
A fast terminal coding agent in Rust for running many AI coding agents in parallel.
- Stars
- 21.4k
- Forks
- 2.5k
- Licence
- MIT
- Language
- Rust
- Last update
- 4 days ago
- Type
- CLI
What it is
jcode is a coding agent you run from the terminal. It uses AI models to read, edit, and work on code in a project, and it is built to run several agents at once without using much memory. The README says agents in the same repo can notify each other when files change, and that they can spawn helper agents for parallel tasks.
- Run multiple agents in the same repository, with the server notifying one agent when another edits a file it has read
- Have a main agent split a task into parallel worker agents
- Resume previous sessions from a session list
- View inline images and LaTeX math within the terminal
Great for
- Developers who want several AI coding agents working on one repo at the same time
- People running many agent sessions who are concerned about RAM and startup speed
- Terminal-first developers comfortable with command-line tools
- Users who want to connect their existing Claude or OpenAI Codex accounts via OAuth
Not ideal for
- Non-programmers, since it is a developer tool that edits source code
- Anyone who needs a graphical editor-style interface, because it runs in the terminal (TUI)
- Teams needing a verified security or sandboxing model, which the README does not describe
How to get it
- 1
See the README
Install on macOS and Linux
curl -fsSL https://jcode.sh/install | bash - 2
Run it
- Run the install command in a terminal
- Open a terminal in a code project directory
- Start jcode and sign in to a provider using the OAuth or provider options described at jcode.sh/docs
- Give the agent a task on your code
- 3
Get the most out of it
- Spawn two or more agents in the same repo so the server coordinates file-change notifications
- Let a main agent act as coordinator and spawn worker agents for parallel subtasks
- Use the resume menu to pick up earlier sessions
- Review the token optimization, KV cache, and measurement notes in docs/HARNESS.md
Downloads
Strengths
- The README benchmarks show low RAM use per additional session and fast time to first input, though these are the project's own measurements
- Built-in messaging between agents in a repo, including direct and broadcast messages
- Supports a wide range of model providers and OAuth logins, per the README
Trade-offs
- Performance claims come from the project's own benchmarks and have not been independently verified here
- The README is short on detail about security, permissions, and how edits are reviewed
- Its self-development mode lets the agent modify and reload its own binary, which needs careful oversight
How it compares
| Project | Stars | Type | Level | Updated | Best for |
|---|---|---|---|---|---|
| jcode (this) | 21.4k | CLI | intermediate | 4 days ago | Developers who want several AI coding agents working on one repo at the same time |
| opencode | 224k | AI model / agent | intermediate | 4 days ago | Developers who want an AI agent that works inside their terminal and existing projects |
| pi | 119k | Dev tool | intermediate | 4 days ago | Developers who want a terminal coding agent they can customize with their own extensions or skills |
| oh-my-pi | 37.5k | CLI | intermediate | 4 days ago | Developers who want an AI agent working inside a terminal on their own codebase |
| Codewhale | 41.2k | CLI | intermediate | 4 days ago | Developers who want an AI agent working inside a terminal and a git repository |
| CowAgent | 47.8k | App | intermediate | 4 days ago | People who want a self-hosted personal assistant reachable from their everyday chat apps |
Stats synced from GitHub. The assessment above was drafted with AI from the project's README on 2026-10-29 and may miss recent changes — check the README before relying on a command. MIT licence: free to use, modify and distribute, including commercially, as long as you keep the copyright notice.