feed · watching · stats · help with your tokens
Muvon/octomind
Octomind is an open-source coding agent harness that gives an LLM real access to a repository through MCP tools.
It is built for agent work that keeps going after the human leaves the screen, enforcing policy as code instead of approval popups. One Rust engine with one TOML config serves an interactive CLI, stdin, a background daemon, WebSocket and an ACP agent.
Similar
- opencode-ai/opencode
A powerful AI coding agent. Built for the terminal.
- anomalyco/opencode
The open source coding agent.
- vercel-labs/agent-desktop
agent-desktop gives AI agents computer use on macOS, Windows and Linux by reading the OS accessibility tree instead of screenshots.
- storytold/craft-anything
Craft Anything is an open-source, pure Rust engine that studies software to understand its mechanics and uses AI agents to build new versions.
- KKKKhazix/agent-autoloop
Agent Autoloop lets a command-line agent pick up work in the background and follow it through, asking for a decision only when something is uncertain.
- aaitor/skill-video-studio
A local video-production workspace driven by a Claude Code skill that turns a product walkthrough into a polished demo video.
- jyotipmahes/ML-Interview-Prep
A repository containing Machine Learning Engineer interview preparation materials and from-scratch algorithm implementations.
- postmanlabs/postman-app-support
The public support repository where Postman users file feature requests, integration ideas and bug reports.
sources
Claims below include README evidence. A matching quote or worker agreement does not establish that a claim is correct. Category majority among 2 workers. Version 1.
Octomind is an open-source coding agent harness that gives an LLM real access to a repository through MCP tools.
Octomind is an open-source coding agent harness.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
It is built for agent work that keeps going after the human leaves the screen, enforcing policy as code instead of approval popups.
Octomind is built for the work that keeps going after you leave.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
One Rust engine with one TOML config serves an interactive CLI, stdin, a background daemon, WebSocket and an ACP agent.
Interactive, stdin, daemon, WebSocket, or ACP — one engine, one config.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Deterministic guardrails: pre-call guards, post-result hooks and post-turn validators.
pre-call guards, post-result hooks, post-turn validators
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
It solved 24 of 25 held-out tasks at lower cost than the closest competitor.
It solved 24 of 25 held-out tasks at lower cost than the closest competitor.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Runs packaged specialists (prompt, model, MCP servers, tool permissions) pulled from a tap registry.
pulls a packaged agent (prompt, model, MCP servers, tool permissions)
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Adaptive compression keeps long tasks moving while retaining task knowledge.
keeps long tasks moving while retaining task knowledge and the live exchange.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Lets you pick a model per role and cap spend per request and per session.
pick a model per role, and cap spend per request and per session
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Turns rules and corrections stated in a session into project-scoped lessons.
Rules and corrections you state in a session become lessons
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Acts as an MCP client for stdio and Streamable HTTP servers with OAuth.
stdio and Streamable HTTP servers, with OAuth
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
An optional sandbox locks writes to the working directory via Landlock on Linux or Seatbelt on macOS.
`--sandbox` locks writes to the working directory via Landlock (Linux) or Seatbelt (macOS)
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Loads standard SKILL.md agent skills on demand so context stays small.
Standard `SKILL.md` skills from `.agents/skills/`
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Multi-step workflows, each step with its own model and toolset.
Multi-step pipelines, each step with its own model and toolset
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Can constrain the final response to a JSON Schema.
Pass `--schema schema.json` to constrain the final response to a JSON Schema
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Runs as an interactive CLI, stdin pipe, background daemon, WebSocket server, or ACP agent using a single binary and config.
One Rust binary, any model, every surface. Interactive, stdin, daemon, WebSocket, or ACP — one engine, one config.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Uses adaptive compression to maintain context and task knowledge during long-running sessions without losing information.
Sessions that stay sharp. Adaptive compression keeps long tasks moving while retaining task knowledge and the live exchange.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Demonstrated effectiveness on real-world pull requests, solving 24 of 25 held-out tasks at lower cost than competitors.
Proven on real PRs. It solved 24 of 25 held-out tasks at lower cost than the closest competitor.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Centralizes all configuration including models, tools, roles, guardrails, and budgets into a single TOML file without framework code.
One config. Models, tools, roles, guardrails, budgets: all TOML, no framework code.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Limitation: Reports anonymous usage by default: commands, tools, models, timings, token counts and error kinds.
Octomind reports anonymous usage: which commands, tools, and models are used
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Limitation: Building from source requires protoc.
Building from source requires `protoc`
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Limitation: The first time you run a specialist, it may install its tools and ask for credentials.
The first time you run a specialist, it may install its tools and ask for credentials.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Limitation: Benchmark numbers are published results from July 30, 2026, not a measurement of the current checkout.
These are the published July 30, 2026 results, not a measurement of this checkout.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Limitation: Over budget, a non-interactive run stops with exit code 75.
a non-interactive run stops with exit code 75
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Limitation: Non-interactive runs stop with exit code 75 when budget limits are exceeded, requiring manual intervention for interactive sessions.
Over budget, a terminal session asks y/N; a non-interactive run stops with exit code 75.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Limitation: Initial specialist runs may require manual credential entry and tool installation before proceeding.
The first time you run a specialist, it may install its tools and ask for credentials.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Limitation: Benchmarks are limited to specific languages and task types from merged pull requests, potentially not covering all domains.
Measured on octobench against 25 tasks harvested from merged pull requests across Python, PHP, Rust, C++, and JavaScript.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Limitation: While it claims not to send sensitive data, users must trust the implementation for complete privacy assurance.
It never sends your code, prompts, file paths, tool arguments, or environment values.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Limitation: Building from source requires additional dependencies like protoc, which may complicate installation for some users.
Building from source requires protoc; see Building from source for per-OS prerequisites.
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
Alternative to Opencode
opencode used the same model and endpoint; octomind solved 24 vs 19 at roughly half the cost
Quote matched the README or GitHub description at extraction. Same claim and quote from 1 worker(s). Interpretation not reviewed by a maintainer.
ai-ml · developer-tools · cli