mlx-code | A Coding Agent That Speaks Git Natively
-->
[ 00_mlx_code ]
> A Coding Agent That Speaks Git Natively
✔ Git isn't just a backup. It's a state machine.
✔ Git commits store file changes together with the conversation that produced them.
✔ Parallel worktrees host multiple isolated agents concurrently.
✔ Agents can spawn sub-agents recursively.
# Branch, diff, and resume from any branching paths dynamically.
[ GITHUB_REPO ]
[ EXPLORE_PRIMITIVES ]
[ 01_cli_demo ]
terminal - zsh
[ IDLE ]
~ $<br>uvx --from mlx-code mlc
[ COPY ]<br>[ ▶ DEMO ]
[ 02_why_git ]
Agent<br>Git<br>What It Enables
Execution step changes files<br>commit<br>Inspect changes cleanly with git diff
Alternative approach<br>branch<br>Explore paths non-destructively
Isolated agent sandbox<br>worktree<br>Keep your active checkout clean and safe
Resuming from a past session<br>checkout<br>Rewind agent to any snapshot state
Most agents treat version control as an afterthought.<br>mlx-code maps the reasoning loop directly onto Git mechanisms. Your complete agent state history lives inside git commit history rather than an opaque database.
[ 03_primitives ]
01 // branching logic paths
Fork from any checkpoint
A branch opens a fresh interactive tab in an isolated worktree.<br>Test divergent code designs securely without abandoning your active checkout.
/branch --rev 2 make it use httpx instead
02 // session recovery
Resume from anywhere
Every file-changing step is snapshot-committed with the chat/execution history.<br>Rewind cleanly to any checkpoints.
mlc --resume abc1234
03 // mid-session hot swaps
Switch configs on the fly
Swap agent configurations mid-session between planner, coder, and reviewer roles without losing your active context.
/clear --config reviewer.yaml
04 // parallel batching
Concurrent multi-agent
Run multiple agents concurrently inside one batch step.<br>Share your unified memory architecture compute pools dynamically.
mlc --engine batch
[ 04_architecture ]
Worktrees:
main ──●──●──●──●──●──●──●──●──●──●──●──●──●──●──────► Node = git commit + chat history<br>│ │<br>│ └── branch-1 ──●──●──●<br>│ │ ┌────────────┐<br>│ └─┤ branch-1-0 ├──●──●<br>│ └─────┬──────┘<br>└── branch-0 ──●──●──● │<br>Tabs: ├────────────► Tab = git branch + Agent<br>┌──────────────────────────────────────────────┼─────────┐<br>│ TUI tabs │ │<br>│ ┌──────┐ ┌──────────┐ ┌──────────┐ ┌─────┴──────┐ │<br>│ │ main │ │ branch-0 │ │ branch-1 │ │ branch-1-0 │ │<br>│ └──────┘ └────┬─────┘ └──────────┘ └────────────┘ │<br>└─────────────────┼──────────────────────────────────────┘<br>Agents: ├──────────────────────────────────────► Each tab runs its own Agent<br>┌────┴─────────────────────────────────────┐<br>│ Agent │<br>│ ┌────────────────┐ ┌────────────────┐ │<br>│ │ API: │ │ Tools: │ │<br>│ │ Local (mlx-lm) │ │ Read Write │ │<br>│ │ Gemini │ │ Edit Bash │ │<br>│ │ Claude │ │ Grep Find │ │<br>│ │ Codex │ │ Ls Skill │ │<br>│ │ DeepSeek │ │ Agent ─────────┼──┼─► Recursively spawns sub-Agents<br>│ └────────────────┘ └────────────────┘ │<br>│ Git worktree │<br>│ (isolation + session state) │<br>└──────────────────────────────────────────┘
[ 05_library_api ]
Use as a Library
Construct custom workflow pipelines, run programmatic background review checks, or coordinate distributed multi-agent workflows natively on your Mac.
from mlx_code.repl import Agent
from mlx_code.tools import ReadTool, WriteTool, EditTool
agent = Agent(api='claude', tool_names=['Read', 'Write', 'Edit'])
result = await agent.run('refactor utils.py to use dataclasses')
[ 06_user_manual ]
slash commands
/branch [--rev N]Fork tab from checkpoint
/diff [--all]Review worktree changes
/clear [--config F]Clear / reset session
/history [--raw]Show chat log
/toolsList active tools
!commandRun a shell command
$commandRun an interactive command
keyboard shortcuts
Ctrl-1 … Ctrl-9Jump directly to N-th tab
Ctrl-, / Ctrl-.Cycle adjacent tabs
EnterSubmit
Ctrl-JInsert newline
Ctrl-RRecall last prompt
Ctrl-CClear input or abort agent
Ctrl-DClose branch tab, or exit app
-->