What is ML.ai
ML.ai is an AI coding agent available for VS Code, Cursor, desktop, and pull requests. It reads your repository, routes each step of a task to the most cost-efficient model, and shows every diff before anything changes. Nothing writes to disk until you allow it. The agent is designed to complete software work end to end—reproducing bugs, editing code, running tests, and opening pull requests—without token budgets or credit meters.
How to use ML.ai
- Install or access: Download the desktop app for macOS, Linux, or Windows, use the IDE extension for VS Code or JetBrains, or install the CLI via
curl -fsSL https://get.mlai.dev | sh.
- Describe a task: Point the agent at a GitHub issue, describe a bug fix, test suite, or migration, or ask a question about your repo.
- Let it work: The agent reads files, searches your codebase, reproduces issues, edits code, and runs your test suite. It routes each step to the appropriate model automatically.
- Review the diff: Before anything is written to disk, ML.ai shows you the diff. Nothing changes until you click Allow.
- Approve the PR: Once tests pass, the agent opens a pull request with a plain-language summary. You review and merge.
Features of ML.ai
- Auto model routing: Each step of a task is routed to the model that fits it, from fast passes on boilerplate to heavyweight models on complex race conditions.
- Diff-first workflow: Every change is shown as a diff before it is applied. Nothing writes to disk until you Allow.
- Parallel agents: Run multiple jobs at once across projects. Each agent works in its own sandbox and reports back with a reviewable diff.
- End-to-end task completion: Reproduces bugs, edits code, runs your test suite, and opens pull requests with plain-language summaries.
- MLAI.md and Skills: Teach the agent your team's standards once with an MLAI.md file and reusable Skills. It applies them on every task.
- Scheduled routines: Schedule routine work like issue triage, dependency audits, and CI watch. The agent runs on its own and pings you only when a decision is yours.
- No token meter: One flat price with no visible token counter. The agent keeps going until the job is verified, not until a meter runs dry.
- Multi-surface availability: Use the same agent in the desktop and web app, inside VS Code or JetBrains, from the terminal, and on GitHub pull requests.
- Stack compatibility: Works with Git, GitHub, Docker, pnpm, pytest, Rust/Cargo, Terraform, Kubernetes, AWS, Prisma, Playwright, and GNU Make.
Use Cases of ML.ai
- Bug fixes: Point it at an issue. It reproduces the bug, edits the code, runs your test suite, and opens a pull request. Example: fixing a double-charge on a checkout page by adding an idempotency key on the client and enforcing it server-side.
- Migrations: Move login off session cookies to passkeys without locking anyone out, adding WebAuthn register and assert while keeping a one-release password fallback.
- Data migrations: Add a
reserved_at column, backfill from stock movements, and ship rollback SQL in the same PR using a @migrate-db skill.
- Routine CI work: Nightly CI watch catches flaky tests, quarantines the spec, replaces a
sleep(500) with waitFor, and opens a patch before standup.
- Parallel project work: Run many jobs at once across projects, such as fixing a double-charge, migrating auth to passkeys, and writing tests for payments simultaneously.
- Background automation: Schedule issue triage, dependency audits, and CI watch to run on their own, pinging you only when a decision is yours.
Pricing
ML.ai Infinite is included in every plan. Pricing is based on concurrency, not tokens.
Individual plans (private-beta launch pricing in India, founder price locked for 12 months):
- Infinite: ₹699/month. Full agent in app, IDE, and CLI; 2 jobs in parallel; auto model routing; no visible token budget.
- Infinite Pro: ₹1,499/month. Everything in Infinite; 6 jobs in parallel; highest-effort models on every step; scheduled routines and GitHub app.
- Infinite Max: ₹3,999/month. Everything in Pro; 20 jobs in parallel; priority queue at peak hours; early access to new models.
Team & Enterprise plans:
- Team: $29/seat/month, billed annually, 5-seat minimum. Infinite Pro for every seat; shared Skills and routines; central usage and approval controls; SSO and audit log.
- Enterprise: Custom pricing, annual contract. Everything in Team; bring your own model endpoints; data residency and private routing; dedicated solutions engineer.
Fair-use and concurrency rules apply. Prices shown are private-beta launch pricing in India and a working hypothesis until general availability; final launch terms, taxes, and regional pricing will be announced at GA. Team pricing shown in USD.
FAQ
What does "Infinite" actually mean?
Infinite means one flat price with no visible token counter. The agent keeps going until the job is verified, not until a meter runs dry. Fair-use and concurrency rules apply.
What is ML.ai Infinite?
ML.ai Infinite is the coding agent that never runs out. It reads your repo, routes each step of a task to the model that fits it, and shows you the diff before anything changes. Nothing writes to disk until you Allow.
Which models does it use?
The agent routes each step to the model that fits it, from a fast pass on boilerplate to a heavyweight model on a gnarly race condition. You set effort and constraints, not vendors. The flagship model inside Infinite's routing pool is P3 Flagship.
How is my code kept private?
Nothing writes to disk until you Allow. Every change is shown as a diff before it is applied. Enterprise plans offer self-hosted or VPC deployment, data residency, and private routing.
Does it work with my existing tools?
Yes. It runs the stack you already have, including Git, GitHub, Docker, pnpm, pytest, Rust/Cargo, Terraform, Kubernetes, AWS, Prisma, Playwright, and GNU Make. It is available in VS Code, JetBrains, desktop, web, terminal, and GitHub pull requests.
Can I review before anything merges?
Yes. The agent opens a pull request with a plain-language summary of every change. You review and merge. Nothing merges on its own.
Is the founder price really locked?
Founder pricing is locked for 12 months on individual plans. Prices shown are private-beta launch pricing in India and a working hypothesis until general availability; final launch terms, taxes, and regional pricing will be announced at GA.
How do I get access?
ML.ai is in private beta, India first. You can download the app, try it in your IDE, or book a call to get started.
