Back to vendors
O

OpenCode

Visit site
Entry priceFree and open source with your own provider keys. OpenCode Go is $10 a month, or Go Plus $40 a month, for a curated model lineup. OpenCode Zen is pay as you go per million tokens, and Enterprise is per seat through sales.Full pricing detail

Open source, model agnostic coding agent from Anomaly, running in the terminal, desktop, web, IDE or as a headless server, with Plan and Build modes and support for 75+ LLM providers using your own keys.

OpenCode is an open source, model agnostic AI coding agent from Anomaly, the team behind SST. It runs in the terminal as a native TUI, and also as a desktop app, a web interface, an IDE extension, a scriptable CLI and a headless HTTP server. It is MIT licensed with about 220,000 GitHub stars and 950 contributors, and OpenCode reports more than 16 million developers using it monthly.

The defining choice is provider independence. Any LLM can be configured with the customer's own API key across more than 75 providers via Models.dev, including local models. OpenCode Zen offers a curated, tested gateway billed per token, and the OpenCode Go plans cover a curated lineup of open and closed models for a flat monthly fee. Work follows a two phase pattern, where Plan mode disables the agent's ability to change anything so it proposes an approach for review, then a single key switches to Build mode to implement it. Multiple agents can run in parallel on the same project, and /undo and /redo traverse change history repeatedly.

OpenCode 2, released on 12 October 2026, runs one background server on the developer's machine that the terminal, web and desktop apps all share. Subagents can work in the background while the main agent keeps going, Code Mode lets an agent chain several tool calls in one script, and more than 80 features ship as internal plugins that hot reload. OpenCode 2 no longer runs language servers, leaning on the project's own lint and typecheck commands instead, and version 1 plugins must be ported. OpenCode 1 stays on its own release line until a developer upgrades.

Project context is grounded in an AGENTS.md file that /init generates from repository analysis and the team commits to git, supplemented by rules, references, @ file mentions and images dropped into the terminal. In OpenCode 1, LSP servers load automatically so the model gets language level intelligence, not just text matching. Sessions persist locally and can be resumed, forked and exported. Sharing sessions by link was discontinued on 15 October 2026.

For organizations, OpenCode Enterprise offers a central config integrating with the customer's SSO provider and routing all traffic through an internal AI gateway with other providers disabled. The architecture stores no code or context data, with all processing local or direct to the chosen provider. The Trust Center shows SOC 2 Type 1 and Type 2 for Anomaly Innovations, Inc., with the report shared under NDA. Extensibility runs through MCP servers, custom tools, plugins, Agent Skills, ACP and a published SDK.

From our comparison video

Crush vs OpenCode: Language Servers or Background Subagents?

Our 2 minute video puts OpenCode side by side with Charm's Crush. It covers what each one does and how each is priced, then when to choose which.

Recorded in October 2026. The profile here is live and updates as research is verified.

Vendor details

Canonical URL

https://opencode.ai

Category

Coding agent

Subcategory

Coding — CLI agent (OSS)

Funding status

Independent, built by Anomaly, the team behind SST, led by Dax Raad and Adam Doty. MIT licensed open source at github.com/anomalyco/opencode with about 220,000 GitHub stars and 950 contributors, reported by the vendor as used by more than 16 million developers a month. Commercial model is OpenCode Zen, a curated hosted model service, OpenCode Go monthly plans, and per seat OpenCode Enterprise.

Company status

independent

Use cases & customers

Primary use cases

agentic codingcode editing

Target customers

developers

Deployment options

Local install (terminal TUI)CLI including scriptable non-interactive modeDesktop appWeb interfaceIDE extensionHeadless HTTP serverDockerEnterprise central config with internal AI gateway

Integrations

Built in tools for file read, edit, search and shell execution, extended by MCP servers, custom tools, plugins and configurable code formatters. LSP servers load automatically to give the model language level code intelligence. Native GitHub and GitLab integrations, ACP support, a headless HTTP server with an OpenAPI spec and an official JS/TS SDK. More than 75 LLM providers are reachable through Models.dev including local models, and GitHub Copilot ships a formal integration authenticating with OpenCode.

In practice

You want a coding agent in your terminal that isn't tied to one vendor's model. OpenCode is model-agnostic, supporting more than 75 providers with your own API key.

Closed coding tools mean trusting a black box with your code. OpenCode is open source, so you can run and inspect it yourself.

You'd rather pay providers directly than a markup. OpenCode is free to run as an open-source tool, with cost only on the model you choose.

Agentic Index coverage score

10.5 / 14 capabilities · 75%

Integrations & Tool Calling Full

Built in tools let the agent read, edit and search files and run shell commands. MCP servers, custom tools, plugins and configurable code formatters extend them, and there are native GitHub and GitLab integrations. OpenCode 1 loads LSP servers automatically for language level intelligence. OpenCode 2, released on 12 October 2026, keeps LSP settings but no longer runs language servers, and points developers to their project's own lint, typecheck and compiler commands instead. Terminal, web and desktop apps sit alongside the IDE extension.

Sourceopencode.ai/docs/tools, docs/mcp-servers, docs/github, docs/gitlab and opencode.ai/v2/docs/migrate-v1read 2026-10-22

Workflow Orchestration Full

A two phase Plan then Build workflow lets the agent propose an implementation approach for review before it makes any changes. Multiple agents can run in parallel on the same project, and custom agents can be configured with their own tools and models. In OpenCode 2, subagents can run in the background while the main agent keeps working, a running subagent or shell command can be moved to the background with one key, and delegated commands report back to the parent session. Custom commands and Agent Skills package repeatable multi step work.

Sourceopencode.ai/docs/agents, docs/commands, docs/skills and anoma.ly/notes/opencode-2-0read 2026-10-22

Knowledge Grounding & RAG Partial

Context is assembled for each run. Built in tools read, grep, glob and list files on demand, and @ fuzzy file references and dropped images attach material to a prompt. OpenCode 1 adds language level diagnostics and symbols from LSP servers at run time, which OpenCode 2 no longer runs. There is no maintained retrieval structure over the customer's code or documents, and the docs have no index or knowledge page. AGENTS.md from /init and rules are instructions the product applies.

Sourceopencode.ai/docs/tools, docs/rules and opencode.ai/v2/docs/migrate-v1; opencode.ai/docs/lspread 2026-10-22

Human Oversight & Guardrails Full

Plan mode turns off the agent's ability to make changes, so it proposes an approach for review before building. A single key toggles it, and the user switches back to Build mode once the plan is agreed. Permissions and policies govern what the agent may do. /undo and /redo revert or replay changes and can be run repeatedly. Non interactive CLI mode auto approves only when it is explicitly invoked for scripting.

Sourceopencode.ai/docs/permissions, docs/policies and docsread 2026-08-30

Security, Identity & Governance Full

Enterprise deployment supports SSO through a central organization config and routes traffic through an internal AI gateway that can switch off every other provider. Permission rules, organization policies and private npm registry support come with it.

The Trust Center linked from the enterprise page shows SOC 2 Type 1 and Type 2 for Anomaly Innovations, Inc., the company behind OpenCode, with the report shared under NDA.

The architecture stores no code or context data, processing is local or goes direct to the customer's chosen provider, and session sharing was discontinued on 15 October 2026.

Sourceopencode.ai/enterprise, opencode.ai/docs/enterprise and trust.opencode.airead 2026-10-25

Observability & Auditability Partial

Sessions are kept in a local database that the CLI can list, query and export as JSON, with a sanitize option that redacts sensitive transcript and file data. opencode stats reports token use and cost by session, model and tool, the TUI shows tool calls as they run, and log levels can be raised for debugging. The agent is MIT licensed, so its behavior is fully inspectable. There is no audit log or organization level record of agent actions.

Sourceopencode.ai/docs/cliread 2026-10-22

Memory & State Persistence Partial

Sessions persist locally and can be listed, resumed, forked, exported as JSON and imported, and /undo and /redo move through change history within a session, so a run carries on where it stopped. Sharing sessions by link was discontinued on 15 October 2026. There is no memory layer that the agent writes and reads across sessions, and the docs have no memory page. AGENTS.md carries project conventions as instructions, not memory the agent keeps.

Sourceopencode.ai/docs/cli, docs/rules and the v1.19.1 release notesread 2026-10-22

Deployment & Data Residency Full

The agent is MIT licensed open source and runs entirely on the developer's own machine. It installs via script, npm, Homebrew, Arch, Chocolatey, Scoop, Mise or Docker. In OpenCode 2 a single background server runs on that machine, and the terminal, web and desktop apps all connect to it. OpenCode stores no code or context data, and all processing is local or goes direct to the customer's chosen provider. Enterprise deployment routes everything through an internal AI gateway with other providers disabled. Session sharing, once the only hosted component, was discontinued on 15 October 2026.

Sourceopencode.ai/docs/enterprise, docs and anoma.ly/notes/opencode-2-0read 2026-10-22

Prebuilt Agents, Templates & Packs Partial

Agent Skills package reusable agent capabilities, custom agents and custom commands can be configured and shared, and plugins extend the runtime. An ecosystem page collects community extensions. There is no catalog of prebuilt agents from OpenCode that install in one click.

Sourceopencode.ai/docs/skills, docs/agents, docs/commands and docs/ecosystemread 2026-08-30

Triggers & Channel Coverage Full

The GitHub integration starts runs without anyone invoking them. The pull_request event runs OpenCode when a pull request is opened, synchronized or reopened, the issues event runs it when an issue is created or edited, and the schedule event runs it on a cron schedule with a stored prompt. A schedule example is published. Comment mentions (/opencode or /oc) and workflow_dispatch cover invoked runs. Beyond GitHub, invocation spans the TUI, a scriptable non interactive CLI returning text or JSON, the desktop app, the IDE extension, a headless HTTP server, ACP clients and a GitLab integration.

Sourceopencode.ai/docs/github (Supported Events, Schedule Example) and /docs/gitlabread 2026-09-29

Model Flexibility & Routing Full

OpenCode is model agnostic by design. Any LLM provider can be configured with the customer's own API key, across more than 75 providers via Models.dev, including local models. There is a provider directory, per agent model configuration, an optional curated OpenCode Zen list, logins with a Claude Pro or Max, ChatGPT Plus or Pro, or GitHub Copilot subscription, and enterprise routing through an internal AI gateway.

Sourceopencode.ai/docs/providers and docs/modelsread 2026-10-22

APIs, SDKs & MCP Extensibility Full

There is a published SDK, a headless HTTP server, a plugin system, custom tools, MCP server support, ACP support and an ecosystem page.

OpenCode 2 rebuilt the server API, added an @opencode/client package and an @opencode/sdk package that runs the same code as the CLI, and its Code Mode turns MCP servers and OpenAPI endpoints into functions an agent script can call several at a time. More than 80 of its features ship as internal plugins that hot reload, and version 1 plugins must be ported.

The whole agent is MIT licensed on GitHub under anomalyco/opencode, and GitHub Copilot ships a formal integration that authenticates with OpenCode.

Sourceopencode.ai/docs/sdk, docs/server, docs/plugins and anoma.ly/notes/opencode-2-0read 2026-10-22

Testing, Debugging & Optimization Partial

OpenCode 1 surfaces LSP diagnostics so the agent can see and correct errors in its own edits. OpenCode 2 drops those diagnostics and relies on the project's own lint, typecheck and compiler commands, which the agent runs through its shell tooling along with tests. /undo and /redo help it iterate on a failed attempt. There is no evaluation harness, replay or regression testing of agent behavior.

Sourceopencode.ai/docs/lsp, docs/tools and opencode.ai/v2/docs/migrate-v1read 2026-10-22

Browser & Computer Use Not documented

The agent works through file operations, shell commands and MCP tools in the terminal, plus LSP servers in OpenCode 1. It has no browser control and cannot operate software that lacks a programmatic interface.

Sourceopencode.ai/docs/toolsread 2026-10-22

The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded

Product updates

·Workflow orchestrationVerified

Version 2.3.0 of OpenCode makes subagents run in the background by default: a subagent call returns immediately, keeps working, and reports back to the parent session when it goes idle. Several subagents can run concurrently while the user keeps working in the parent session, and a subagent with its own unfinished background work now tells the parent it is still running so the parent waits for the final report. The release also adds experimental support for MCP events behind the mcp_events flag, and lets model providers present a client certificate to gateways that require mutual TLS.

Bears on: Workflow orchestration

View source
·Workflow orchestrationVerified

Worktree support reaches scripted runs in OpenCode v2.2.6, so the opencode run command can start a headless or CI session in a fresh git worktree, named or not. MCP servers can also be pooled, letting projects with the same server setup share one connection, which means fewer processes and faster startup.

Bears on: Workflow orchestration

View source
·Agent capabilityPartially Verified

OpenCode v2.2.1 adds a /goal command in the terminal app and Desktop that keeps a session working toward a stated objective until a reviewer confirms it is met. Progress is reviewed automatically along the way, and the user can pause, resume or clear the goal at any point. Plugins also get an experimental hook to carry on work before a session goes idle.

Bears on: Workflow orchestration

View source
View all 7 OpenCode updates →Tracked since Jun 2026 · Verified from public vendor sources

Pricing

Free and open source with your own provider keys. OpenCode Go is $10 a month, or Go Plus $40 a month, for a curated model lineup. OpenCode Zen is pay as you go per million tokens, and Enterprise is per seat through sales.

Hybrid. The agent is free with the customer's own keys, OpenCode Go is a flat monthly plan with per model allowances, OpenCode Zen bills per million tokens against a prepaid balance, and Enterprise is per seat.

Free tier

Included quota

Free with your own keys (BYOK) includes the full OpenCode feature set with any provider or local model and no included quota, since the customer funds its own tokens. Go at $10 a month covers a curated model lineup, open and closed, with a monthly usage allowance set in dollar value and drawn down within 5 hour, weekly and monthly windows.

What is public

The core agent is free and MIT licensed, used with the customer's own keys across 75+ providers or local models. OpenCode Go has two monthly plans, Go at $10 and Go Plus at $40, each with monthly allowances and per million token prices published per model. OpenCode Zen is pay as you go with per million token prices for every listed model, including free models offered for a limited time, auto reload of $20 when the balance drops below $5, and workspace and per member monthly limits. OpenCode Enterprise is priced per seat through sales, with no token charge when the customer uses its own LLM gateway.

Billing mechanics

The tool itself is free, and cost comes from model access by one of three optional paths. The customer can bring its own key and pay the provider directly. Zen is a curated gateway billed per million tokens against a prepaid balance, with optional auto reload and monthly limits. Go is a flat $10 or $40 monthly plan with per model allowances that can fall back to the Zen balance when a limit is reached. Enterprise adds a per seat license.

Cost watchouts

Token spend on the customer's own key is uncapped. Zen auto reload can exceed a monthly limit that is set, and Go falls back to the Zen balance only when that option is enabled.

Variable cost rationale

Model usage is the metered cost. On the free core the customer pays its own provider per token, with nothing from OpenCode in between, and Zen bills per million tokens against a prepaid balance. Either bill grows with how much the agent runs and which models it calls. Go is the limit. Its flat $10 or $40 a month carries a usage allowance capped by dollar value, and when the allowance runs out usage stops unless fallback to the Zen balance is switched on. Zen and Go let workspaces and members set monthly spend limits, though Zen's auto reload of $20 can charge past a limit that is set. Usage on the customer's own key has no cap from OpenCode, and Enterprise is a per seat license with no token charge on the customer's own gateway.

Additional watchouts

The core is free with the customer's own key. Go covers a curated model lineup only, and Go and Go Plus are for a single member workspace that cannot invite members. Zen auto reload can charge past a monthly limit that is set.

Overage / add-ons

Go usage is capped by dollar value. When the allowance runs out, usage stops or falls back to a Zen pay as you go balance. Zen reloads $20 automatically when the balance drops below $5, and the reload can be configured or switched off. Usage on the customer's own key bills directly from the provider with no cap.

Sales call required

No, self serve available

Free / trial

Free and open source agent, used with the customer's own provider keys or local models

Lowest paid plan

OpenCode Go $10 a month

Commercial notes

The agent is MIT licensed. Go has two tiers ($10 and $40), and Enterprise is per seat, with no token charge on a customer's own gateway.

Key ambiguities

Go's monthly allowances are set per model, so what a plan covers depends on model choice, and the Enterprise per seat rate is not published.

Cancellation / refund

No long term contract, and plans cancel at any time. Zen runs on a prepaid balance, and Zen and Go let workspaces and members set monthly spend limits.

Support SLA / resale

Community support on the free open source core, and managed support through Zen, Go and Enterprise.

Missing data

The Enterprise per seat rate is not published.

Agentic Index verified 2026-10-25

Alternatives to OpenCode

The closest documented capability profiles to OpenCode among coding agents tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.

  • Poolside11.0 / 14Fuller documented coverage on Observability & Auditability
  • Ellipsis10.5 / 14Fuller documented coverage on Observability & Auditability
  • Grok Build10.5 / 14Fuller documented coverage on Prebuilt Agents, Templates & Packs
  • Tabnine11.5 / 14Fuller documented coverage on Knowledge Grounding & RAG and Observability & AuditabilityOpenCode vs Tabnine →
  • Amp12.0 / 14Adds documented Browser & Computer Use
  • Augment Code12.0 / 14Fuller documented coverage on Knowledge Grounding & RAG and Observability & Auditability

Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded

Contact us

Found a vendor we missed? Have feedback on the index? We'd love to hear from you.