OpenCode
OpenCode is an MIT-licensed coding agent available as a terminal interface, beta desktop application, and IDE extension. It works inside a project, where agents can inspect files, propose plans, edit code, and run commands. The product emphasizes provider choice rather than requiring one model vendor, and its documentation covers built-in agents, permissions, tools, LSP integration, MCP servers, configuration, session sharing, GitHub, and GitLab workflows. Beta desktop downloads are published for macOS, Windows x64, and Linux, while the CLI also supports package managers, an install script, Docker, and release binaries.
The OpenCode software is open source. Users can connect their own model-provider API keys, including local or OpenAI-compatible options. OpenCode Zen is optional: its official docs describe a curated model gateway with account credits, pay-as-you-go requests, and a changing set of free and paid models. OpenCode inference is not uniformly free: Zen carries a changing mix of free and paid models, while external providers bill under their own terms.
Top features
- Three working surfaces: terminal TUI, beta desktop app, and IDE extension provide separate interfaces for project work.
- Repository actions: agents can search and read a project, edit files, and execute commands with configurable permissions.
- Provider flexibility: connect supported hosted providers, local models, or OpenAI-compatible endpoints.
- LSP integration: language-server diagnostics and project intelligence can inform code changes.
- MCP support: local and remote MCP servers can add tools or data sources to an agent session.
- Built-in modes: documented Build and Plan agents separate full tool access from a read-oriented planning workflow.
- Cross-platform distribution: official options cover macOS, Linux, Windows, Docker, npm-compatible package managers, Homebrew, Chocolatey, Scoop, and release binaries.
Use cases
- Diagnose a failing test from the terminal, edit the relevant files, and rerun the check.
- Use the beta desktop app to organize several coding sessions while reviewing project changes.
- Work over SSH, in containers, or on developer machines where model-provider choice is required.
- Create a read-oriented implementation plan before granting an agent permission to modify files.
- Connect internal documentation or developer tools through MCP for repository-specific tasks.
No linked agents yet.
Register an agentGrok Build 1.0 Is xAI’s Bid to Make Terminal Coding Agents Usable Every Day
Grok Build 1.0.0 arrived on August 7, 2026. Its headline is not a new model or benchmark score, but the less glamorous work of making an agent that edits code, runs commands, and keeps working easier to control and recover.
AI Tools · Aug 8, 2026
Cloudflare Built a Browser for Agents, Not Humans, and Put It on Workers
Kitesurf is a Chromium alternative for agent workloads: Rust/Wasm pieces, V8 isolates, CDP compatibility, 3–7× less CPU/memory on Cloudflare’s corpus, and an honest list of things it still cannot do.
ai-agents · Aug 7, 2026
ChatGPT Puts GPT-5.6 Sol on Instant, Unlimited Luna on Free, and Leaves Work and Codex Alone
OpenAI’s August 6, 2026 ChatGPT update unifies paid Instant and reasoning on a retuned Sol, moves Free and Go to Luna with a later unlimited-text window, and keeps agent surfaces on the July models.
ai-models · Aug 6, 2026

