Pricing
Pay for token usage, CPU, memory, and platform fees. New organizations get started with a $100 credit. No per-seat fees or idle charges.
Engineers shipping code to your repositories
Pull requests opened per developer each week
Share of those PRs written by @ellipsis
Review every PR against your rules and past feedback
3 PRs/day for a team of 10 developers using AI for 50% of PRs
Simple metering for each session. Nothing idle, no seats.
| Ellipsis Cloud | BYOK Bedrock | Private Cloud | |
|---|---|---|---|
| Tokens | At cost, no markup | On your AWS bill | $0 |
| CPU | $0.142 / vCPU-hour | $0.142 / vCPU-hour | $0 |
| Memory | $0.024 / GB-hour | $0.024 / GB-hour | $0 |
| Platform fee | 100% of token usage | 100% of token usage | 100% of token usage |
BYOK Bedrock routes model tokens through your own AWS Bedrock account, so they land on your AWS bill and you can spend down your AWS credits. CPU and memory still run on Ellipsis compute.
Metering
Examples
The same per-session prices the estimator uses, broken down by what each session meters.
Ellipsis reads the diff, checks it against your team's conventions and past review feedback, then leaves inline comments. No sandbox spins up, so you pay only for tokens and the fee.
Someone shipped a query without an "is not null" guard and prod is throwing. Ellipsis reads the Sentry stack trace, finds the unguarded access, adds the check, and opens a PR before the team wakes up.
A minor version bump changes a function signature. Ellipsis updates every call site, runs your test suite until it is green, and opens the PR.
Coverage dropped on a new endpoint. Ellipsis writes a test that exercises the happy path and the edge cases, confirms it passes, and pushes it.
A test passes locally but fails one CI run in ten. Ellipsis reproduces the failure, traces the race across files, and pushes a fix.
Add an API route with a schema, validation, and tests. Ellipsis spins up a sandbox, writes it end to end, runs the tests, and opens the PR.
Every new organization starts with a one-time $100 credit ($10 for individual accounts). It applies to your usage automatically, and covers roughly 135 code reviews or 50 small features before you pay anything.
Every agent session runs in its own sandbox, and you pay for exactly what that session consumes: CPU and memory metered by the second (rates above), tokens passed through at Claude's rates with no markup, and a per-session platform fee equal to 100% of the session's token usage. Nothing runs idle, so nothing bills idle.
A quick fix, such as a 2am Sentry alert, lands near $0.69 total. A code review on a medium PR is about $0.74. A flaky test investigation is about $1.48. Bigger tasks with more files and longer sessions cost more.
No. There are no per-seat fees. Your entire team can use Ellipsis. You only pay for what your agents do.
We support open source projects case by case. If you maintain one, email team@ellipsis.dev and we can set you up with free usage.
Yes. Set monthly spending caps per team or per repo in the dashboard. Once the cap is reached, agents pause until the next billing cycle.
Yes. Volume discounts, priority support, dedicated onboarding, and custom SLAs. Email team@ellipsis.dev to discuss.