
Amazon Q Developer
Generative AI coding assistant for building, operating, and transforming software
Discover top open-source software, updated regularly with real-world adoption signals.

Terminal‑native AI coding assistant, provider‑agnostic and extensible.
A fast, open‑source AI coding agent that runs in the terminal, supports any LLM provider, and offers a client/server TUI for local or remote workflows.

OpenCode is an AI‑powered coding agent designed for developers who live in the terminal. It works with any large language model—Anthropic, OpenAI, Google, or self‑hosted alternatives—so you stay free from vendor lock‑in while getting contextual code suggestions.
The tool provides a full‑screen TUI built by neovim enthusiasts, and its client‑server architecture lets the backend run on a workstation while you interact from another terminal or a mobile device. Installation is a single curl script or via npm, Homebrew, or Arch packages, with the install path configurable through environment variables.
Released under the MIT license, OpenCode encourages contributions and custom extensions. Whether you need quick snippet generation, remote assistance, or integration into CI pipelines, the agent adapts to a variety of terminal‑centric workflows.
When teams consider OpenCode, these hosted platforms usually appear on the same shortlist.
Looking for a hosted option? These are the services engineering teams benchmark against before choosing open source.
In‑terminal code generation
Generate functions or snippets directly while editing in neovim, reducing context switches.
Remote AI assistance from mobile
Run the server locally and control it via a mobile client to get suggestions on the go.
Provider‑agnostic experimentation
Swap between Anthropic, OpenAI, or local models without changing the client, enabling cost and performance testing.
Automated code review in CI
Integrate the CLI into CI pipelines to have the AI suggest improvements or catch simple bugs.
OpenCode is fully open‑source, not tied to a single provider, emphasizes a terminal UI, and uses a client‑server model that allows remote control.
Any model that offers an API compatible with the OpenAI format works, including Anthropic, OpenAI, Google, or self‑hosted local models.
Yes. The server component can run on any machine, and you can connect via the TUI from another terminal or a mobile client.
The installer respects $OPENCODE_INSTALL_DIR first, then $XDG_BIN_DIR, $HOME/bin, and finally $HOME/.opencode/bin as fallbacks.
Currently the primary interface is the terminal UI; a mobile client exists for remote control, but no full desktop GUI is provided.
Project at a glance
ActiveLast synced 4 days ago