OpenCode alternatives in 2026: 8 coding agents and apps to use instead

Disclosure: Enjoy is our product, and it runs OpenCode, Claude Code and Codex. We checked every option below against its maker’s own pages on November 12, 2026, and we say where another agent is the better pick, OpenCode included.

OpenCode alternatives compared

OpenCode, from Anomaly, “is an open source coding agent,” available “as a terminal-based interface, desktop app, or web app,” per its docs. It’s free, under the MIT license, with more than 226,000 stars on GitHub, and its homepage says it’s “Trusted by 22M+ developers.” These are the agents and apps people weigh when they want something else, side by side.

Option	Best for	Price, checked November 12, 2026	Runs on	Models
OpenCode	Choosing the model, free and open source	Free; Go $10 a month; Go Plus $40; other providers at their prices	Mac, Windows, Linux; desktop and web apps; iPhone in TestFlight	200+ providers, ChatGPT and Copilot plans, free and local models; Claude by the token
Claude Code	The Claude plan you already pay for	Claude Pro $20 a month, or $17 billed yearly; Max $100; Ultra $200	Mac, Windows, Linux	Anthropic’s Claude models
Codex CLI	Your ChatGPT plan, in an open-source agent	ChatGPT Plus $20 a month; Pro $100, $200 or $500; or an API key	Mac, Windows, Linux	OpenAI’s models; others through its config, local ones with --oss
GitHub Copilot CLI	Claude, GPT and Gemini on one plan	Free; Pro $10 a month; Pro+ $39; Max $100	Mac, Linux; Windows in PowerShell or WSL	Claude, GPT and Gemini; Opus 5.5 and GPT-6.1 Sol from Pro+; or your own provider
Kilo CLI	Staying closest to OpenCode	Free; models at provider rates through its gateway; Kilo Pass from $19 a month	Mac, Windows, Linux; also VS Code and JetBrains	500+ models, or your own keys
Cline	Approving every edit, in an editor or a terminal	Free for individuals; ClinePass $9.99 a month	VS Code, JetBrains IDEs, a CLI; a desktop app for Mac and Windows	Any provider’s key; open-weight models on ClinePass
Pi	A minimal agent you shape yourself	Free; models at their providers’ prices	Mac, Windows, Linux	Anthropic, OpenAI, Google, OpenCode Zen and Go, and many more
Crush	Asking before every tool call, in any terminal	Free; models at their providers’ prices	Mac, Windows, Linux, Android, FreeBSD, OpenBSD, NetBSD	Any OpenAI- or Anthropic-compatible provider, Charm’s Hyper, and local models
Enjoy	Claude Code, Codex and OpenCode in one conversation	A monthly subscription, free for 7 days, with your phone and one computer (pricing)	Mac, Windows, Linux	Claude Code, Codex, Grok Build, Antigravity, Muse Code, OpenCode

Prices are each maker’s US list prices per month, before tax. The free agents run on your own plans and API keys, which those companies bill for separately.

Why people look for an OpenCode alternative

OpenCode is popular for good reasons, but four things send people looking, each in OpenCode’s own words:

• Your Claude plan doesn’t work in it. Of plugins that bring Claude Pro or Max into OpenCode, its provider docs say “Anthropic explicitly prohibits this,” and they stopped coming bundled “as of 1.3.0.” Anthropic’s terms say it “does not permit third-party developers to offer Claude.ai login into their own applications, or to route requests through Free, Pro, Max, or Ultra plan credentials on behalf of their users.” So Claude in OpenCode bills by the token: Opus 5.5 is $4 per million input tokens and $20 per million output on its Console.
• It asks little before it acts. “Every agent, including custom agents, starts with this ordered base policy,” says its permissions page, and the policy’s first rule allows every action. Only reading .env files asks first.
• Its free models are temporary. Its Console docs say they’re free “for a limited time,” and for several, “The team is using this time to collect feedback and improve the model.”
• Its Console was breached. On October 17, OpenCode said “We had a security breach last night” that “impacted 1.4% of OpenCode Console users,” “a subset of users who had external providers linked.” Our memo covers which keys to replace.

Anthropic’s rule isn’t about OpenCode alone. The same terms say they don’t prevent “an end user from signing in to the unmodified Claude Code binary with their own Claude subscription,” so for a Claude plan, the agent to use is Claude Code itself, on its own or in an app that runs it.

Claude Code

Best for: using the Claude plan you already pay for.

Claude Code is Anthropic’s own coding agent, and the only agent here that runs on a Claude plan. Its setup guide lists macOS 13.0 and later, Windows 10 1809 and later, and Ubuntu, Debian and Alpine Linux, and there’s a desktop app for macOS, Windows and Linux if you’d rather not use a terminal. It “requires a Pro, Max, Ultra, Team, Enterprise, or Console account,” and “The free claude.ai plan does not include Claude Code access.”

Price: Claude’s pricing puts Pro at $20 a month, or $17 a month billed yearly; Max at $100, for “5x Pro usage”; and Ultra at $200, with “10x Pro usage” a week and “No 5-hour limit” for verified accounts.

Watch for: “Pro and Max plans have a five-hour session limit and a weekly limit,” says Anthropic’s help center, and “These limits are shared across Claude and Claude Code.” It runs Claude only: Anthropic “doesn’t support routing Claude Code to non-Claude models through any gateway,” per its gateway docs. It isn’t open source: its license reads “© Anthropic PBC. All rights reserved.” And on native Windows, its setup guide lists sandboxing as “Not supported”; WSL 2 supports it. Our OpenCode vs Claude Code guide compares the two in depth.

Codex CLI

Best for: your ChatGPT plan, in an open-source terminal agent.

OpenAI’s Codex CLI is a “Lightweight coding agent that runs in your terminal,” under the Apache-2.0 license, with 135,000 stars on GitHub. It installs on macOS, Linux and Windows, and you “Sign in with ChatGPT,” which OpenAI recommends on Plus, Pro, Business, Edu and Enterprise, or use an API key. It isn’t locked to OpenAI: “A model provider defines how Codex connects to a model,” says its configuration guide, and “Codex can run against a local ‘open source’ provider such as Ollama or LM Studio when you pass --oss.”

Price: Codex’s pricing gives Plus, at $20 a month, “GPT-6.1 Sol and GPT-6 Luna” “on the web, in the CLI, in the IDE extension, and on iOS.” Pro is “Plans at $100, $200, or $500 USD per month,” and “Pro plans currently have no five-hour limit.” Free and Go, at $8, get “GPT-6 Luna at Standard speed in the desktop app, subject to rollout.”

Watch for: limits aren’t fixed numbers. OpenAI’s message estimates “are not fixed message limits,” and “Weekly limits may also apply.” A ChatGPT plan pays for OpenAI’s models only, so another provider’s models, set up in its config, bill at that provider’s prices.

GitHub Copilot CLI

Best for: Claude, GPT and Gemini on one flat plan.

“The command-line interface (CLI) for GitHub Copilot allows you to use Copilot directly from your terminal,” says GitHub. It runs on Linux and macOS, and on Windows “from within Powershell and Windows Subsystem for Linux (WSL).” You change models with /model, and “You can configure Copilot CLI to use your own model provider instead of GitHub-hosted models,” “including locally running models such as Ollama.”

Price: “All plans include Copilot CLI and the Copilot app,” says its plans page. Free has “Auto model selection only.” Pro is $10 a month with 1,500 AI credits, and “A selection of models,” among them Claude Sonnet 5.5 and GPT-6 Luna. Pro+, at $39 with 7,000 credits, and Max, at $100 with 20,000, add Claude Opus 5.5, GPT-6.1 Sol, and Gemini 4 Argon and Carbon.

Watch for: “AI credits are consumed based on the number of tokens processed,” so long contexts and high reasoning use them faster. With options such as --allow-all-tools, Copilot can act with your own file access, and GitHub warns that “Scoping of permissions is heuristic.”

Kilo CLI

Best for: staying closest to OpenCode, with a gateway that sells models at cost.

“Kilo CLI is a fork of OpenCode,” says the FAQ on Kilo’s GitHub, where it’s under the MIT license with 28,000 stars. Its CLI page promises “Any model, zero markup: frontier, open weight, or your own keys. One agent, fully open source,” with 500+ models and “Full BYOK support.” It publishes builds for Windows, macOS and Linux, and Kilo has a VS Code extension and a JetBrains plugin too.

Price: Kilo’s pricing starts free. Its gateway says “Pay exact provider rates with no markup.” Kilo Pass, a monthly subscription to credits, is Starter at $19 a month for “Up to $26.60/mo in credits,” Pro at $49 and Expert at $199. Teams is $15 a user a month.

Watch for: a “Credit purchase processing fee” of 5% on every tier. Kilo says its CLI is “Not a thin wrapper,” but as a fork it’s the smallest step away from OpenCode, so if OpenCode’s design is what you’re leaving, look further down this list.

Cline

Best for: approving every edit and command, in your editor or a terminal.

Cline is an “Autonomous coding agent as an SDK, IDE extension, or CLI assistant,” under Apache 2.0, with 71,000 stars on GitHub. It comes as a VS Code extension, a JetBrains plugin, a CLI, and “a native app for macOS and Windows,” and “Cline is not locked to a single AI provider.” Where OpenCode allows everything by default, its Plan and Act modes work the other way: “Every file edit and terminal command requires your approval, so you stay in control of what actually changes,” unless you turn on auto-approve.

Price: Cline’s pricing calls it “free for individual developers,” with “Model Inference at Cost or BYOK.” ClinePass, its answer to OpenCode Go, is “$9.99/month with generous quotas” for open-weight models such as GLM 5.3, Kimi K3, DeepSeek V4 Pro and Qwen3.8-Max.

Watch for: ClinePass includes “capable open weight models,” not Claude or GPT; for a wider list, Cline’s own provider charges “pay-as-you-go API rate,” or you bring a key. Its desktop app isn’t on Linux.

Pi

Best for: a minimal agent you shape yourself, with any provider.

“Pi is a minimal agent harness,” says its site: “There are many agent harnesses but this one is yours.” It’s under the MIT license, with 121,000 stars on GitHub, and installs on macOS, Linux and Windows. It works with “Anthropic, OpenAI, Google, Azure, Bedrock, Mistral, Groq, Cerebras, xAI, Hugging Face” and more; “run /login inside Pi to connect a subscription or API key.” Its providers page takes a key for “OpenCode Zen and Go,” so an OpenCode Go plan comes with you.

Price: free; models cost what their providers charge.

Watch for: it’s minimal on purpose. “Pi ships with powerful defaults but skips features like sub-agents and plan mode,” and “Pi does not include a built-in permission system for restricting filesystem, process, network, or credential access.” Its README suggests a container or sandbox if you need stronger boundaries.

Crush

Best for: an agent that asks before every tool call, in any terminal.

Crush, from Charm, is “Glamourous agentic coding for all,” with “first-class support in every terminal on macOS, Linux, Windows (PowerShell and WSL), Android, FreeBSD, OpenBSD, and NetBSD,” per its GitHub, where it has 29,000 stars. It takes any provider with an OpenAI- or Anthropic-compatible API, and local models through Ollama, llama.cpp and LM Studio. “By default, Crush will ask you for permission before running tool calls.” Charm’s Hyper “is the official model provider for Crush,” “built on open-weight models,” with “subscription and prepaid options.”

Price: free; models cost what their providers charge.

Watch for: it isn’t open source in the usual sense yet. Its license, FSL-1.1-MIT, allows “any purpose other than a Competing Use,” and each version becomes MIT “on the second anniversary of the date we make the Software available.” With Hyper, “your Crush usage metrics may be associated with your Hyper account.”

Enjoy

Best for: Claude Code on your Claude plan, Codex on your ChatGPT plan and OpenCode for everything else, in one conversation.

Enjoy is a desktop app for Mac, Windows and Linux that runs Claude Code, Codex, Grok Build, Antigravity, Muse Code and OpenCode, on the AI plans you already have. It installs OpenCode for you and lists the models OpenCode offers from your providers, local ones included. What it adds for someone leaving OpenCode:

• Your Claude plan, through Claude Code. Enjoy runs Claude Code itself, signed in to your Claude account, so the plan you pay for does the hard work while OpenCode keeps the cheap and local models.
• One conversation, any agent. Switch agents at any point: “The new agent gets the conversation so far, including replies from other agents, before it works on your message.”
• A way through limits. When an agent runs out, Enjoy switches to another of your accounts with usage left, or waits for the reset and carries on. Select Stop waiting to retry with OpenCode instead.
• OpenCode’s approvals, as cards. “When OpenCode’s permission settings say to ask first, Enjoy shows an approval card in the conversation.”
• A night shift. A scheduled workflow can run OpenCode with a local model overnight, with no plan to run out.

Price: Enjoy is one monthly subscription, and the first seven days are free, with a card to start. It includes your phone, any browser and one computer; each other computer costs extra each month, and each teammate you invite needs a subscription of their own. See pricing.

Watch for: Enjoy doesn’t bring your Claude plan into OpenCode: Claude Code runs on it, and OpenCode on the providers you add with opencode auth login. Enjoy shows Usage unavailable for OpenCode, “because usage depends on each provider.” It runs six agents, so Kilo, Cline, Pi, Crush and Copilot CLI aren’t among them. Scheduled workflows run only while your computer is awake with Enjoy open. Enjoy isn’t open source, and it’s a monthly subscription on top of your AI plans once the free week ends.

When to stay with OpenCode

Stay if you like choosing the model. OpenCode signs in to ChatGPT Plus or Pro and GitHub Copilot, takes “200+ LLM providers through Models.dev, including local models,” and costs nothing itself, with OpenCode Go at $10 a month if you want one flat price across its 32 models.

If the permissions are what bother you, you don’t need to leave: OpenCode’s rules take an "ask" effect, so you can make edits and commands wait for you. If it’s Claude, add Claude Code beside it rather than replacing it, as our OpenCode vs Claude Code guide describes, and if it’s a desktop app you want, our OpenCode GUI guide compares the apps that run it.

How we compared

We read each maker’s homepage, pricing, docs, license and GitHub repository on November 12, 2026, and every page is linked below. Star counts come from GitHub that day. We quote each maker’s own words for its limits, and we leave out what a maker hasn’t documented. Agents and plans change often, so we recheck this guide against its sources at least once a month, and the date at the foot of the page says when we last did.