LogoAI Nexus
Ante logo

Ante

Visit

Antigma Rust harness. Rechecked: 1,846 stars. Public repo Apache-2.0; binary free in preview. CLI $0 plus BYOK or local GGUF. Not a $20 seat.

Share:

Ante

Ante is Antigma Labs' self-contained terminal coding agent. Rechecked 2026-08-25: github.com/AntigmaLabs/ante had 1,846 stars. Source in that repo (docs, SDK, protocol crates) is Apache-2.0. The prebuilt ante binary is a private-core preview: BINARY-TERMS.md says it is free to run, including commercially, during the preview. The CLI itself is $0. Models are BYOK, subscription OAuth, or a local GGUF via built-in llama.cpp. Docs say macOS and Linux only; Windows users are pointed at WSL. First-party intro post is dated 2026-03-31. Do not invent a dollar seat.

Compare Claude Code if you wanted Anthropic's CLI, OpenCode if you wanted another open harness, and Crush if you wanted Charm's TUI plus Hyper.

Key Features

  • One Rust binary: install with curl -fsSL https://ante.run/install.sh | bash. README says ~15MB compressed, zero runtime deps. Core harness is still private; this public repo is docs, protocol, and eval adapters.
  • TUI, headless, serve, gateway: ante, ante -p, ante serve, ante gateway (Slack/Discord).
  • Providers: Anthropic, OpenAI, Gemini, Grok, DeepSeek, OpenRouter, local GGUF, plus ~/.ante/catalog.json for custom OpenAI-compatible endpoints. No Antigma account required.
  • Yolo flag: ante --yolo skips tool approval prompts. --permission-mode yolo is the same class. Docs warn destructive tools normally need approval.
  • Profiles: --profile pi or bare swaps the whole settings file. Skills, MCP, and memory are first-party.

Limitation: preview software. Breaking changes are expected. The 7x / 9x / 5x footprint claims versus Claude Code are first-party Docker numbers on docs.antigma.ai, not an independent audit. Terminal-Bench 2.1 82.7% is a pinned Ante 0.preview.71 run on DeepSeek V4 Flash, not a forever score.

Use Cases

  • People who want a Claude Code-shaped loop without Node, and will attach BYOK or a GGUF.
  • People who will stay offline with ante --offline-model path/to.gguf.
  • People comparing Rust harnesses before paying Claude Code.

Pricing

Binary is free in preview. Model access on 2026-08-25.

Path Price Notes from first-party pages
Public repo $0 Apache-2.0. 1,846 stars on the check date.
ante binary $0 in preview Commercial use allowed under BINARY-TERMS. Terms may change after preview.
BYOK / OAuth Provider bill Anthropic, OpenAI, Gemini, Grok, DeepSeek, OpenRouter, and catalog entries.
Local GGUF $0 inference Built-in llama.cpp. Hardware is yours.
Antix Console Recheck live Optional account/API dashboard at portal.antigma.ai. No public dollar seat on the pages we opened.

If a leftover blog prints Ante Pro $20, treat this table as the source.

Getting Started

  1. macOS or Linux (or WSL). Run curl -fsSL https://ante.run/install.sh | bash.
  2. Start ante. Pick a provider, or /offline-mode for a local GGUF.
  3. Headless: ante -p "find and fix the failing test".
  4. Do not pass --yolo outside a disposable sandbox.

First-party start: docs.antigma.ai/start/quickstart and the GitHub README.

Frequently Asked Questions

Is the binary open source?

Not the core harness. The public repo is Apache-2.0 docs/SDK. The binary is preview terms.

Official on Windows?

Not native. First-party docs say use WSL.

Same as Claude Code?

No. Ante is Antigma's harness. Claude Code is Anthropic's.

Alternatives

Tips

  1. Quote 1,846 stars and the preview binary terms. Do not call the core MIT.
  2. Recheck antigma.ai/eval before you paste 82.7%.
  3. Skip --yolo on a machine you do not trust.

Conclusion

Ante is a 1,846-star Rust terminal harness with a free preview binary and BYOK or local GGUF, not a leftover $20 seat. Start at github.com/AntigmaLabs/ante, then decide whether Claude Code already covers the workflow you need.

Comments

No comments yet. Be the first to comment!

Related Tools

Autohand Code logo

Autohand Code

www.autohand.ai/code

Self-evolving terminal agent. Rechecked: Apache-2.0 with ARR cap, 178 stars. CLI $0 plus BYOK. Enterprise is sales, not a public $20 seat.

Dev Tools
Continue logo

Continue

continue.dev

Open-source coding agent for VS Code and JetBrains. Apache-2.0, local models, CLI, and MCP. Acquired by Cursor; the repo stays public.

Dev Tools
Mistral Vibe logo

Mistral Vibe

github.com/mistralai/mistral-vibe

Mistral terminal agent. Rechecked: Apache-2.0, 4,872 stars. CLI $0 plus Mistral API key. --agent auto-approve / --yolo. Not a $20 seat.

Dev Tools

Related Insights

What Locks Codex Is the Picker, Not the Models

What Locks Codex Is the Picker, Not the Models

You already pay for OpenCode Go, Grok, and Z.ai, but the Codex picker still shows mostly GPT. The community project codex-router does not teach another install ritual. It puts subscriptions you already bought back into the selector, keeps keys on the machine, and leaves native GPT alone.

Hook OpenCode Go into Codex on Windows. Do Not Open a Second Toolkit.

Hook OpenCode Go into Codex on Windows. Do Not Open a Second Toolkit.

A ChatGPT-signed Codex desktop app still shows mostly GPT in the picker. On Windows, enable only OpenCode Go and the Grok, GLM, Kimi, DeepSeek, and MiniMax models you already pay for appear in the same selector. Keys stay local. Native GPT stays put.

Seven AI Coding CLIs, Six Months: No Matter How Strong the Model, Work Needs Supervision

Seven AI Coding CLIs, Six Months: No Matter How Strong the Model, Work Needs Supervision

Claude Code, Codex, opencode, pi, omp and DeepSeek Harness all have personalities. After six months of deep use I run a division of labor: pi for the fastest cheapest reviews, omp for complex PRs, DeepSeek Harness on V4 Flash for high-frequency low-cost review, and Claude Code, Qoder and Cursor for writing. No matter how strong the model, work needs supervision — ideally from an independent third party.