What Is OpenCode?Start Here
The open-source terminal AI agent explained: how the MIT-licensed tool works, the bring-your-own-key model across 75+ providers, its build and plan agents, and why the software is free while the model cost is yours.
405 W. Greenlawn Ave Lansing, Michigan 48910
contact@techjacksolutions.com
+1-616-320-4064
AI Coding Assistant
An open-source (MIT) AI coding agent from Anomaly that runs in your terminal. It is privacy-first, brings your own API key across 75+ providers and local models, understands your code through language servers, and ships build and plan agents. The software is free; you pay your chosen model provider directly.
Terminal TUI
Switch agents with Tab, drag-drop images, and use /undo and /redo in your shell
Bring Your Own Key
Connect 75+ providers via Models.dev, or run local models with Ollama and LM Studio
Build & Plan Agents
A full-access build agent and a read-only plan agent, plus custom agents you define
LSP-Enabled
Auto-loads language servers so the agent understands your code, not just text
Desktop Beta, IDE Extension & MCP
A desktop app in beta and an editor extension complement the terminal, with Model Context Protocol support
OpenCode is an open-source AI coding agent built by Anomaly and released under the MIT license. It runs primarily as a terminal TUI, with a desktop app in beta and an IDE extension. It is privacy-first and brings your own model: you connect your provider keys, and the software itself stays free. The figures below labeled vendor-reported come from OpenCode's own materials, verified 2026-06-16.
OpenCode is MIT-licensed and developed by Anomaly. It does not store your code or context data, which makes it appealing for teams that want an auditable, self-directed agent. The project reports roughly 175k GitHub stars and around 940 contributors (vendor-reported, 2026-06-16). It can also read existing Claude Code config files such as CLAUDE.md and skills, unless you disable that behavior.
OpenCode is bring-your-own-API-key through the AI SDK and Models.dev, supporting 75+ providers such as Anthropic, OpenAI, DeepSeek, Google Vertex, Bedrock, and Groq, plus local models via Ollama, LM Studio, and llama.cpp. Because there is no inherent model fee, every dollar of model spend is your provider's cost, not a flat subscription. You can also link an existing ChatGPT, GitHub Copilot, or GitLab Duo plan at no extra charge.
A default build agent has full access, a plan agent is read-only, and you can author custom agents. OpenCode is LSP-enabled so it auto-loads language servers, runs multiple sessions in parallel, and supports MCP. Optional paid layers exist, OpenCode Zen for curated coding models pay-as-you-go and OpenCode Go as a low-cost subscription, but neither is required to use the tool.
Vendor-reported, verified 2026-06-16. OpenCode is open-source and BYO-key, so the software is free and your only cost is the model provider you choose. Confirm current details at opencode.ai and opencode.ai/docs.
MIT
Open-Source License
$0
Software / BYO Key
75+
Model Providers
940+
Contributors (Reported)
In-depth coverage of what OpenCode is, how to install and use it, and how it compares with Cursor. Researched with vendor documentation and honest trade-offs.
The open-source terminal AI agent explained: how the MIT-licensed tool works, the bring-your-own-key model across 75+ providers, its build and plan agents, and why the software is free while the model cost is yours.
A walkthrough from install to first results: setting up the terminal TUI, connecting a provider key or a local model, switching between the build and plan agents, and reviewing edits before you keep them.
The open-source terminal agent against the agent-native editor: OpenCode's bring-your-own-key freedom and privacy-first design versus Cursor's polished IDE experience, with an honest read on where each fits.
Explore the model providers you can plug into OpenCode, a rival AI code editor, and the broader AI Tools Hub.
Cursor Hub
The agent-native AI code editor compared head to head with OpenCode.
Anthropic Claude Hub
The Claude models you can connect as a provider inside OpenCode.
ChatGPT Hub
OpenAI's GPT models, another provider you can bring your own key for in OpenCode.
AI Tools Hub
Breakdowns, comparisons, and guides across the leading AI vendors.
AI Governance
Responsible AI, code provenance, the EU AI Act, and compliance for generative tools.
Important context for responsible AI adoption
OpenCode positions itself as privacy-first and states that it does not store your code or context data. In practice, your prompts and code are still sent to whichever model provider you connect, so the privacy you get depends on that provider's terms. Running a local model with Ollama or LM Studio can keep code on your machine, while a hosted provider processes it on their infrastructure. Review the data-handling terms of each provider before processing sensitive or proprietary code, and prefer local or enterprise-controlled providers for confidential work.
Terminal coding agents like OpenCode speed up output, but long agent-driven sessions can blur into overwork, and the ease of generating code should not replace rest, review, or real connection. If you are experiencing distress:
AI systems can produce plausible-sounding but incorrect guidance. For mental health, medical, legal, or financial decisions, always consult a qualified professional.
See the NIST AI Risk Management Framework for structured guidance on AI risk assessment.
Under GDPR (EU) and CCPA (California), you have the right to access, correct, and delete your personal data, though for OpenCode those rights are exercised mainly through whichever model provider processes your requests. AI coding tools can produce confident but wrong output, so always review agent edits before keeping them and validate generated code against your own tests and security review.
The EU AI Act sets transparency obligations for generative AI, including disclosure of AI-generated content. Organizations adopting AI coding tools remain responsible for meeting these provisions and for the code they ship.
This publication is editorially independent. AI tool coverage reflects independent research, vendor documentation, and editorial judgment. Where affiliate links are present, they are clearly disclosed and do not influence conclusions.