- Tool type
- Terminal coding agent
- Made for
- Developers using terminal and CI
- Runs on
- Local terminal on macOS, Linux, and Windows
- Installed as a native muse binary; Windows uses PowerShell and lacks voice input and session messaging.
What it does
Muse Code is a coding agent you run from the command line in a project directory. Given a task, it plans, edits files, and runs commands, with approvals and an OS sandbox enabled from the first run. It takes a natural-language prompt plus your workspace (skills, rules, and hooks load when you trust the workspace) and produces code changes, command output, and a session you can review. A headless mode runs a single prompt to completion for scripts and CI.
People use Muse Code for multi-file coding in an interactive terminal session, or in pipelines via muse exec. Distinct features include workflows that coordinate parallel subagents (where available), session messaging between live local sessions, rewind to earlier conversation points, and a TypeScript SDK in developer preview that drives the same engine. It runs on macOS, Linux, and Windows, with PowerShell replacing Bash on Windows and voice input and session messaging not available on Windows.
It suits developers who want a ready-made coding agent at the command line rather than building their own on the API. Sign-in uses a browser or an API key, and subscription plans start at $5/month, with usage limits per 5-hour window. Key limitations: workflows depend on build and rollout, Windows lacks voice and session messaging, and some capabilities vary by plan.
What it can do
- Plans, edits files, and runs shell commands to complete coding tasks from a prompt
- Interactive terminal session with slash commands, approvals, and live status
- Headless mode (muse exec) to run a single prompt to completion for scripts and CI
- Workflows that coordinate parallel and staged subagents when the installed build supports them
- Session messaging that lets live local sessions exchange handoffs and status updates (macOS and Linux)
- Rewind a conversation to an earlier safe point while preserving workspace files
Example workflows
- Multi-file feature change in an interactive session
- Inputs: a project directory and a prompt describing the feature. Steps: run muse in the project, trust the workspace, describe the change, review proposed edits and approve commands. Outputs: edited files and command results shown in the session.
- Headless run in CI
- Inputs: a single prompt and META_API_KEY set in the environment. Steps: run muse exec with the prompt in the pipeline. Outputs: the agent completes the prompt non-interactively and returns its result to the job log.
Before you choose
- Subscription plans are monthly with usage limits per 5-hour window; the product page lists $5, $15, $50, and $200 per month tiers with different usage multiples.
- Workflows, session messaging, and some features depend on the installed build and rollout, and are not available on every platform.
- Muse Code is a separate product from the Model API; pay-as-you-go API rates apply when calling the API directly, and the Model API pricing page governs those rates.
Listed under
Get started
Sign in via browser or set META_API_KEY for non-interactive environments.
- Install with the one-line installer (curl -fsSL https://dev.meta.ai/install.sh | bash on macOS/Linux, or the PowerShell installer on Windows) and confirm with muse --version.
- Run muse in a project directory and trust the workspace on first entry.
- Authenticate with a browser sign-in or an API key; for CI, set META_API_KEY.
Plans and prices
Prices in USDMonthly subscription tiers with increasing usage multiples. Prompt limits are stated per 5-hour window for the Everyday plan. Model API pay-as-you-go rates are separate and are governed by the Model API pricing page; the product page also lists per-million-token rates for muse-spark-1.3 and muse-spark-1.3-contributor.
- Access to Muse Spark, Muse Wonder and Muse Image
- Send 10-50 prompts every 5 hours, including image and video uploads
- 5x more usage than the Everyday Usage plan
- More multimodal inputs
- Access to the most powerful Muse models, including Muse Wonder Full
- 20x more usage than the Everyday Usage plan
Source: dev.meta.ai/products/muse-code, checked 28 Oct 2026. Prices can change; check before you buy.
Sources for this overview
Product pages, documentation, and pricing checked 28 Oct 2026. Each source supports part of the overview; availability and prices can change.
- Muse Code | Meta ↗dev.meta.ai
- Muse Code - Meta Model API ↗dev.meta.ai
- Muse Code: New plans and features ↗dev.meta.ai
- Get started with Meta Model API and Muse Code - Muse Wonder, Muse Spark, Muse Image, Muse Voice Transcribe, Segment Anything Model, Muse Spark Open, and Muse Glimmer - Meta Model API ↗dev.meta.ai
Ratings and reviews
No ratings for Muse Code yet.
What’s changed
Free accountSignals
Free accountData and security
Free accountCompare with similar tools
Free accountWhere this comes from
How we research- Company website
- dev.meta.ai: product, pricing and policy pages
- YC directory mirror
- Batch and company name
- Web search
- Confirms the official site
- Social accounts
- Accounts the company links, and its posts on X
- Human review
- Checked by a person
- knowledge_base
Last checked 28 Oct 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.