Goose by Block
Open-source local AI agent, Block-built, Linux Foundation-run
Overview
Goose is an Apache-2.0 open-source AI agent created at Block and donated to the Linux Foundation's Agentic AI Foundation in December 2025. It runs locally, works with 15+ LLM providers including local Ollama models, and extends via MCP.
Pricing
Pricing shown for reference only. These figures reflect RECATOOLS research as of 11 Jul 2026 and may be out of date or incomplete. This is not financial or purchasing advice — always confirm the current price on the provider’s official website before making any decision.
Use cases
ASEAN Perspective
Goose by Block in Southeast Asia
ASEAN-region availability and pricing notes coming soon. Drop the editorial team a note via /contact/ if you can supply local context (Singapore/Malaysia/Indonesia/Thailand/Vietnam).
Goose is a free, Apache-2.0 AI agent that started at Block and moved to the Linux Foundation's Agentic AI Foundation in December 2025, alongside MCP and AGENTS.md. It runs the full dev loop on your own machine — editing files, installing packages, running tests — with any of 15+ LLM providers including local models via Ollama, and extends through MCP servers. For teams that want no vendor lock-in and local execution, it's one of the strongest open options; development is brisk (v1.41 shipped July 2026) and the community is large. The trade-offs are real: you bring your own API keys, output quality tracks whichever model you plug in, and there's no commercial support. Self-hosting suits data-residency needs across ASEAN. Excellent value, with a developer-grade learning curve.
What people say
Fifty-one thousand GitHub stars, a v1.41 release in July 2026, and a governance handoff to the Linux Foundation — Goose has quietly become the establishment choice among open-source coding agents. Block built it internally and donated it in December 2025 to the new Agentic AI Foundation, where it sits alongside Anthropic's Model Context Protocol and OpenAI's AGENTS.md as an inaugural project. For teams nervous about betting on a single vendor's agent, that neutral governance is the whole pitch.
Technically it's a Rust binary with a desktop app and a CLI that runs the full dev loop on your machine: edit files, install packages, execute shell commands, run tests. You bring the model — 15+ providers are supported, from Anthropic and OpenAI down to local Ollama models — and extend it through MCP servers, of which the ecosystem now counts 70-plus.
User reports are consistent on both the good and the bad. The good: genuine model freedom (swap in a cheaper model for grunt work), offline and local operation for data-residency-sensitive teams, and zero licence cost under Apache 2.0. The bad: output quality is only as good as the model you plug in, BYO-key setups mean watching your own token bills, and the polish trails paid rivals — several 2026 reviews still frame it as better than you'd expect but rougher than Claude Code. There's no commercial support line; you file GitHub issues like everyone else. For tinkerers and platform teams, that's a feature rather than a bug.
Summary of public user & expert reviews, compiled by RECATOOLS.
About this listing
This entry was compiled from publicly available data including Goose by Block's official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with Goose by Block unless explicitly stated.
Third-party AI tools update their pricing, features, availability, and policies frequently. Information here may be outdated by the time you read this — we make reasonable efforts to keep listings current, but cannot guarantee absolute accuracy.
For the latest details, please refer to Goose by Block directly →
Spotted something out of date? Suggest an update →
Goose by Block in the news
More in Code & Dev Tools
