Skip to content

Ostra

Models trained on how software gets tested, grounded in your app's memory.

The family

One family. A model for each testing job.

Use Ostra Pro where you would use Claude or Codex for testing work. Use Ostra Decision when CI or an agent needs a fast verdict.

Live · 1.0

Ostra Pro

Writes and fixes tests, analyses failures, reasons about a flow, and reads screenshots and video. Use it where you would use Claude or Codex for testing.

$0.64
Input · per 1M tokens
$0.06
Cached input · per 1M tokens
$4.47
Output · per 1M tokens
Explore Ostra Pro
Beta

Ostra Decision

One model that makes three testing calls: why a test failed, which tests a diff affects, and whether a selector will survive the next release.

$0.50
Triage · per 1M tasks
$2.00
Test impact · per 1M tasks
$0.25
Selector survival · per 1M locators
Explore Ostra Decision

Ostra Code

A model for writing and maintaining test code, in the framework and conventions your suite already uses.

Explore Ostra Code

Ostra Visual

Visual regression with high-precision DOM change detection, with support for video.

Explore Ostra Visual

Benchmarks

Priced against the coding models you use today.

What a real testing request costs on Ostra and on other coding models, at published list prices. Each model page has the full comparison.

Ostra Pro: cost to fix failing tests
  • Ostra Pro
  • Other coding models
Failing tests fixed against total cost
Model020406080100
Ostra Pro$0.00$0.82$1.64$2.46$3.28$4.09
Claude Sonnet 5.5$0.00$2.40$4.80$7.20$9.60$12.00
Claude Opus 5.5$0.00$4.80$9.60$14.40$19.20$24.00
GPT-5.6 Terra$0.00$2.48$4.96$7.44$9.92$12.40
Grok 4.7$0.00$2.24$4.48$6.72$8.96$11.20
Ostra Decision: cost to triage failures
  • Ostra Decision
  • Coding models asked the same question
Failures triaged against total cost
Model02004006008001,000
Ostra Decision$0.00$0.0001$0.0002$0.0003$0.0004$0.0005
Ostra Pro$0.00$1.55$3.10$4.64$6.19$7.74
Claude Sonnet 5.5$0.00$4.60$9.20$13.80$18.40$23.00
Claude Opus 5.5$0.00$9.20$18.40$27.60$36.80$46.00
GPT-5.6 Terra$0.00$4.72$9.44$14.16$18.88$23.60
Grok 4.7$0.00$4.36$8.72$13.08$17.44$21.80

Surfaces

Everywhere you work

One agent across every surface.

Studio

Watch an agent read a failing run, form a plan, and work the fix — every action visible before it lands.

Open Studio
TestRelic Studio · shop-e2e4m 08s

Ostra Pro · working the fix

What it did

Diagnosed the runrun #48213 failures
Recalled project memory[data-test=pay-now] renamed in v2.4.13 notes
Proposed a fixcheckout.spec.ts+9 −4
Nothing applied without approvalawaiting you
Re-runs the test after you approveautomatic

TUI

The same agent in your terminal. One Rust binary, artifacts rendered in your scrollback, a full MCP host.

Install the CLI
testrelic — shop-e2e

▌why is checkout flaky on firefox?

Ostra Pro

● Checked how often this spec fails(spec: "checkout.spec.ts")

⎿ 4/30 fails · P75 3.2s → 8.4s

● Diagnosed the failing run(run: "#4821")

⎿ 3 failures · 214 specs

▎Not a timing flake. [data-test=pay-now] was renamed to [data-test=pay-submit] in v2.4.1 — all 4 failures are post-deploy.

MCP

Point Cursor, Claude Code, Copilot or Codex at one server. 39 tools, one token.

Connect an agent
mcp.json · ~/.cursor39 tools

Ostra Pro · testrelic registered

Cursorconnected
Claude Code
Copilot
Codex

Server

Transportstdio or hostednpx @testrelic/mcp
Authone personal tokenscoped
Same memory in every clientorg-wide

SDKs

Playwright, Appium and Maestro — JS or Python. A run becomes memory the moment it finishes.

Get the SDK
shop-e2e · reportersrun #4820

Run uploaded · 214 specs

Playwright
Appium
Maestro
pytest

One line in your config

ReporterJS or Python@testrelic/*
A run becomes memory when it finishesautomatic
Readable by Ask AI, Studio and any MCP clientorg-wide

Console

Create an API key, add prepaid credit and see what every request cost.

Open the Console
TestRelic Console

Ostra · key ready

Balance

Prepaid creditAdded by card, from $5$25.00

API keys

ci-pipelineFull access · $50 capactive
dashboardRead only · expires in 30 daysactive
One key for every Ostra modelPro · Decision

Memory + models

A model that knows testing, on facts about your app.

A general model guesses at your app. Ostra is trained on the work testing teams do, and TestRelic memory supplies what is true about your suite.

Memory says what happened

Runs, flakes, selector changes and your team's decisions, captured by the SDK as every suite finishes.

Ostra decides, or writes

Ostra Decision returns a verdict. Ostra Pro reasons about the failure, writes the fix and reads the screenshots.

The result goes back to memory

Verdicts and fixes are visible to Ask AI, Studio, the CLI and every MCP client on the team.

Pricing

Prepaid. No plan needed to try it.

Add credit, create a key and make a request. You pay for what you use, from one balance.

  1. Add credit

    Pay by card, adding anything from $5 to $999 at a time. Turn on auto top-up and the balance refills itself when it falls below a level you set.

  2. Create a key

    Full access or read only, with an expiry, a spend cap and an IP allowlist. One key works for every model and every surface.

  3. Pay per use

    Each request is billed at the model's rate per 1M, as listed on the pricing page. Every model draws on the same balance.

How it is billed

Included in plans, or on its own
Growth, Pro and Enterprise include Ostra usage. Without a plan, Ostra bills from its own prepaid balance.
A limit on spend
Set a monthly spend limit for the account, with a webhook alert, and a cap on each key.
Never billed past your balance
Ostra only runs a request your balance can cover. If it cannot, the request is refused, and you add credit to carry on.

Frequently asked questions

Ostra is TestRelic's family of models for testing work, trained on how software gets tested and grounded in your app's memory. Ostra Pro and Ostra Decision are available today; Ostra Code and Ostra Visual are coming soon.

Ostra Pro, which is live, is an advanced model for testing software: it writes and fixes tests, analyses failures, and reads screenshots and video. Ostra Decision, which is in beta, returns a verdict on a failure, a diff or a selector.

Use Ostra Pro where you would use Claude or Codex for testing work: when you need the reasoning, the fix, or a test written. Use Ostra Decision when CI or an agent needs a fast verdict: why a test failed, which tests a diff affects, or whether a selector will survive. Ostra Decision returns a verdict, not an explanation, so send the failure to Ostra Pro for that.

In Studio, in the CLI, over MCP from Cursor, Claude Code, Copilot or Codex, and through the REST API. All four use the same API key and the same prepaid balance, which you manage in the Console.

Sign in to the Console with your TestRelic login, add your billing details, add credit from $5, and create a key. Then use the key in Studio, the CLI, over MCP or through the REST API.

Ostra Pro is $0.64 per 1M input tokens, $0.06 per 1M cached input tokens and $4.47 per 1M output tokens. Ostra Decision is $0.50 per 1M tasks for triage, $2.00 per 1M tasks for test impact, $0.25 per 1M locators for selector survival. Billing is pay as you go from a prepaid balance, and top-ups start at $5. Ostra usage is also included in Growth, Pro and Enterprise.

No. You can pay as you go from a prepaid balance, with no plan. Growth, Pro and Enterprise also include Ostra usage, which is metered and resets automatically: Pro has 10x Growth's, and Enterprise is unlimited.

A request with no balance is refused until you add credit. You can add $5 to $999 per payment. To limit spend, each key can carry a spend cap, and the account has a monthly spend limit.

TestRelic memory supplies what is true about your suite: runs, flakes, selector changes and your team's decisions, captured by the SDK as every suite finishes. The results go back to memory, and are visible to Ask AI, Studio, the CLI and every MCP client on the team.

The Benchmarks above compare what a real testing request costs on Ostra and on other coding models, at their published list prices, and each model page has the full comparison. We have not published a quality comparison against coding models.

Both are coming soon, and we have not announced a date. Ostra Code is for writing and maintaining test code. Ostra Visual is for visual regression, with support for video. To talk to the team about either, book a call.

One key, four ways in

Put a model that knows testing on your team.

Create a key in the Console, add credit, and call Ostra from wherever you already work.