# AntigmaLabs/ante

Ghost in your shell. Ante is a self-contained agent harness with a highly optimized core. It works like Claude Code or Codex, with none of their dependencies or model constraints.

Repository: https://github.com/AntigmaLabs/ante
Canonical: https://ross.abutalabs.com/products/antigmalabs-ante
Homepage: https://antigma.ai
Language: Rust
License: Apache-2.0
License Family: permissive
Topics: ai-agents, rust, agent, agentic-ai, artifical-life, terminal
Last push: 2026-08-26T16:58:52+00:00

## Health v2 (maintenance only)
Score: 79/100 (v2, computed 2026-09-03T02:39:23.370411+00:00)
- activity 99, release rhythm 87, longevity 18
- inputs: {"age_days": 253, "days_push": 7, "days_rel": 7, "gap_med": 1.0, "n_releases_24m": 83}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1908, forks 59 (observed 2026-08-28T04:05:52.015732+00:00)

## What it is
Ante is a self-contained terminal-based coding agent harness written in Rust, distributed as a single ~15MB binary with zero runtime dependencies. It works with any LLM provider (Anthropic, OpenAI, Gemini, local GGUF models via llama.cpp) and can serve as an optimized core for building custom agent harnesses.

## Use cases
- run a claude code alternative in the terminal
- code with an ai agent fully offline using local models
- build my own agent harness on an optimized core
- orchestrate multiple ai coding agents in parallel
- use an ai coding agent without vendor lock-in or accounts
- run local gguf models for agentic coding
- reduce memory and cpu usage versus claude code

## When to choose
- you want a lightweight, dependency-free terminal coding agent
- you need offline/local model inference with no API keys
- you want to switch freely between 12+ LLM providers
- you're building a custom agent harness and need an optimized core
- you run many agents in parallel and care about resource footprint

## When to avoid
- you need Windows support natively (macOS/Linux only; use WSL)
- you require a stable, production-hardened tool (beta preview with breaking changes)
- you prefer a GUI-based coding assistant
- you want a harness co-trained with a specific model like Claude Code

## Facets
- artifact type: cli-tool
- maturity: experimental
- function: agent-framework, llm-inference, cli, developer-tools, mcp
- domain: large-language-models, developer-tools
- platform: cli, rust
- tags: coding-agent, terminal-agent, local-inference, llamacpp, self-contained-binary, multi-agent-orchestration, offline-mode, claude-code-alternative, ai-agents, command-line, macos, linux

## Member repositories
- AntigmaLabs/ante (main) score 79

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:52.015732+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T03:11:10.584246+00:00, confidence not recorded.
  - readme: https://github.com/AntigmaLabs/ante (fetched 2026-08-28T04:05:52.015732+00:00, sha 1c7867df1c14)
  - homepage: https://antigma.ai (fetched 2026-08-29T10:50:14.302556+00:00, sha 0350cc09d301)
  - site_page: https://antigma.ai/about (fetched 2026-08-29T10:50:14.305313+00:00, sha 2868fac0a1af)
  - site_page: https://docs.antigma.ai (fetched 2026-08-29T10:50:14.307085+00:00, sha 56a006d05ac0)
- Data as of 2026-08-30T08:39:29.467469+00:00.