Free & Open Source
$0
MIT license, BYOK
OpenCode
The open-source agent that plans, builds, and ships — from any client.
The open-source AI coding agent from Anomaly (SST). Provider-agnostic — use Claude, OpenAI, Google, or local models — with a client/server architecture that lets a single headless agent be driven from a TUI, desktop app, or IDE extension. Built-in Build and Plan agents, MCP and LSP integration, ~160K GitHub stars.
+Pros
- Provider-agnostic: works with Claude, OpenAI, Google, Bedrock, OpenRouter, and local models
- Client/server architecture allows one headless agent driven from TUI, desktop, or IDE
- 160K+ GitHub stars — fastest-growing open-source coding agent in 2026
- Build and Plan agents for both execution and read-only analysis modes
- Privacy-first: no code or context stored server-side
−Cons
- Younger project — fewer integrations and polish compared to Cursor or Copilot
- Requires self-managed LLM API keys and provider setup
- TUI interface has a learning curve for non-terminal-native developers
- No built-in hosting or deployment capabilities
Pricing
OpenCode is the open-source AI coding agent from Anomaly — the team behind SST and terminal.shop. Provider-agnostic by design, it works with Claude (recommended), OpenAI, Google, OpenRouter, Bedrock, and local models, and ships with OpenCode Zen: a curated set of models benchmarked for coding agents. By mid-2026 it had grown to roughly 160K GitHub stars, 900 contributors, and around 7.5 million monthly developers.
OpenCode's client/server architecture is the key design choice: the agent runs as a headless server, and the TUI, desktop app, and IDE extensions are interchangeable clients — you can run the server on a workstation and drive it from a laptop or phone. Two built-in primary agents (Build and Plan) let you switch between full-access execution and read-only analysis, with custom subagents defined as simple markdown files. The agent integrates LSP for type-aware reasoning, supports MCP servers, and keeps your code local — OpenCode does not store user code or context.
Key Features
• Client/server architecture: headless server with TUI, desktop, and IDE clients
• Provider-agnostic: Claude, OpenAI, Google, Bedrock, OpenRouter, local models
• OpenCode Zen: curated, benchmarked model set for coding agents
• Build and Plan agents: switch between full-access and read-only modes
• Custom subagents via AGENTS.md and markdown files
• LSP integration: TypeScript, Python, Go, Rust, Java, and more
• MCP support and custom commands
• Multiple concurrent sessions with their own context
• Privacy-first: no code or context stored server-side
Version History
Core: headerTimeout config, experimental background agents push updates without polling, optional modalities config. Bugfixes for MCP disconnect.
Core: redesigned diff viewer with file tree. Bugfixes for model errors, PTY sessions, MCP OAuth config.
Core: Grok OAuth sign-in with device-code login. Improved error handling. PDF attachment support for Grok.
Core: diff viewer in TUI, shell mode for run prompt, Anthropic API-key models use native runtime.
Signature Snippet
$ opencode
› opencode
⋮
build | Add a /webhooks endpoint and tests for it
plan: 1) Add route in src/server/routes.ts
2) Validate payload with zod schema
3) Wire into worker queue
4) Add tests in tests/webhooks.test.ts
▶ Approve plan? Y/nMore in this Space
Prime Agent (Prime Intellect)
Prime Agent is an open-source (MIT) self-improving coding and research harness by Prime Intellect, built on a Recursive Language Model that runs inside a persistent IPython REPL and a Continual Harness that refines its own state via /refine. It took the same frontier weights from 30.2% to 95.5% on ARC-AGI-3 RHAE Best@1, powered daemon-backed sessions, and ships no hosted offering.
OpenCodeReview (ocr)
OpenCodeReview (ocr) is Alibaba's open-source AI code-review CLI. It reads git diffs and drives an LLM agent through a hybrid architecture where deterministic pipelines handle file selection, rule matching, and comment positioning while the agent does the reasoning. Built on two years of internal use at Alibaba scale, it ships a multi-language ruleset and consumes roughly 1/9 of the tokens of general-purpose agents.
Delta (by Zed)
Multiplayer environment from Zed for coding with agents. Threads unify the agent conversation, the code edits, and human review in one shareable artifact, replacing the branch/PR loop. Built on DeltaDB, a CRDT-based version control layer that records every edit between commits alongside the prompts and reasoning that produced them.