The usage model is a deliberate design choice by the ex-Sourcegraph team behind it. There is no cheap model to swap in and no way to cap spend per request. That makes cost control a workflow problem: what you ask, how often you approve, and how long you leave the agent running.

Why is there no free tier?

The team argued that a free plan forces them to route users to a weak model, which produces useless output and a bad first impression. Instead Amp always runs frontier models and charges for them. Reasonable, but it means you cannot experiment for free.

What actually drives the bill?

• Number of agent turns, including retries after failed commands.
• How much of the repository is pulled into context per turn.
• Model choice, which is automatic and not user-configurable.
• Long-running sessions that keep thinking while you do something else.

The retry loop is the hidden cost. If a test fails and the agent keeps rerunning it with variations, every attempt bills. Capping that behaviour is the single biggest lever you have.

How do you scope a refactor to stay cheap?

Break the job into discrete requests with a clear finish line: “rename this type across the repo and update imports” is measurable. “Improve the codebase” will run indefinitely. Review the plan before approving multi-file edits, and stop the session once the diff is correct rather than asking for a further polish pass.

What does a realistic month cost?

These are approximations, not published tiers. The point is the spread. A team that budgets a flat $10 and lets everyone run all day will get a shocking invoice.

How do you forecast before committing?

Run one representative week, log the usage, then multiply. Do not guess from a single afternoon; cargo-culting a low estimate is how teams end up arguing with finance. Watching the first month’s real invoice is the only honest calibration.

Who should skip Amp entirely?

Anyone whose work is mostly in-editor autocomplete. Amp is built for terminal-driven, multi-file agent work, and using it as a fancy tab-completion engine is an expensive way to buy a feature you could get free elsewhere. If your budget is tight and your needs are light, look at a cheaper coding assistant first.

Key Takeaways

• Amp has no free tier; plans start around $10/mo plus usage.
• Retry loops and long sessions are the biggest cost drivers.
• Scope each request to a measurable finish line and stop when done.
• Forecast from a real week of usage, not a single session.

That last point is the whole argument. Amp’s designers decided a free plan would push them toward a weak model and a crippled experience, so they charge from the first token. Plans start around $10 per month and scale with usage, which means your bill depends on how much you let the agent run.

What is Amp and who makes it?

Amp comes from Ampcode, built by engineers who previously worked on Sourcegraph and at Google. It is deliberately opinionated: terminal-first, frontier models only, no manual model switching. The agent decides when to edit files, run shell commands, and search the repo.

How do you install and start Amp?

Install the CLI, authenticate, then run it inside a project directory:

npm install -g @sourcegraph/amp
amp login
cd ~/projects/my-app
amp

Extensions exist for VS Code and JetBrains if you prefer the editor, but the terminal is the main path. Within a minute or two of first use you will see the agent proposing file edits and waiting for approval.

How does Amp pricing actually work?

Costs track the models and tool calls the agent uses. A heavy afternoon of multi-file refactoring can cost more than the base subscription. There is no way to force a cheaper model, which is intentional and also the most common complaint.

What is Amp good at?

• Command-line refactors across many files, driven by one plain-English request.
• Repository chores such as dependency bumps and lint fixes.
• Preparing a branch for review by running tests and summarising failures.
• Navigating unfamiliar legacy code with guided questions.

It is weakest for people who want a cheap autocomplete. If your work is mostly tab-completion in an editor, Amp is an expensive way to get it.

How do you keep Amp costs under control?

Scope each request tightly, review the plan before approving multi-file edits, and avoid leaving long-running agent sessions open while you think. Because billing is usage-based, an agent that keeps retrying a failing command is an agent that keeps spending. Watching the first month’s invoice is the only honest way to calibrate.

Key Takeaways

• Amp is a terminal-first coding agent with no free tier.
• Plans start near $10/mo plus usage-based charges.
• Model routing is automatic; you cannot pick a cheap model.
• Best for shell-centric developers who value opinionated defaults.