Pricing
Credits are dollars.
Pay only for successful answers, from a USD balance. Every plan gets every engine and every feature. No seats, no per-engine fees, no per-domain fees.
Free
Try every feature on a real brand.
- $1 of usage every month
- Every engine, MCP, CLI and REST
- 1 project
- Unlimited free test keys
- No card needed
Pay-as-you-go
Spiky or occasional workloads.
- Top up from $10
- Credits never expire
- Same per-call prices
- Every engine and feature
Builder
One brand, tracked daily.
- $60 of usage every month
- +22% vs pay-as-you-go
- 5 projects
- Unused usage rolls over one month
Pro
Teams tracking a category across engines.
- $260 of usage every month
- +31% vs pay-as-you-go
- 25 projects, 5 seats
- Alerts on significant changes
- Unused usage rolls over one month
Agency
Agencies running many client brands.
- $840 of usage every month
- +40% vs pay-as-you-go
- Unlimited projects
- White-label reports
- Bring your own provider keys
- Budgets per sub-account
Enterprise
Volume commits and procurement.
- Negotiated usage commit
- SSO
- SLA
- Warehouse sync
- Everything in Agency
Plans are prepaid usage with a bonus, not feature gates: every plan gets every engine and every MCP, CLI and API feature. Pay-as-you-go top-ups never expire. Prices in USD, excluding tax.
Estimate
What will tracking cost?
Set your panel and see the monthly cost and the cheapest way to pay for it. Your agent can get an exact quote with dry_run.
Worked examples
Real workloads, real numbers.
Daily brand tracking
One brand, 50 buyer prompts, the two consumer apps, 5 samples, every day, queued.
500 × $0.003 = $1.50/day × 30
≈ $45/month on pay-as-you-go
Category coverage across engines
50 prompts on 4 engines (2 consumer apps queued, Perplexity and Gemini via API), 5 samples, weekly.
2 APIs: 50 × 2 × 5 × $0.05 = $25
$26.50/run × 4.3 runs
≈ $113.95/month → Builder + top-ups
Controlled experiment
Did the new comparison page change Claude's and Grok's answers? 20 prompts × 2 premium engines × 10 samples, before and after.
400 × $0.10 × 2 runs
= $80, with a significance test
Site audit on every deploy
100 pages crawled as AI bots, 5 rendered in a browser, run in CI on each pull request.
= $0.03 per audit
300 deploys ≈ $9/month
Price list
Every call, every price.
The same list is machine-readable at /pricing.json and live at GET /v1/prices.
| Work | Unit | Price | Per $1 |
|---|---|---|---|
| AI answers | |||
| Consumer UI (chatgpt_ui, google_ai_mode), live | answer | $0.01 | 100 |
| Consumer UI, queuedresults in minutes | answer | $0.003 | 333 |
| Native API + web searchchatgpt, gemini, perplexity | answer | $0.05 | 20 |
| Premium model + web searchclaude, grok | answer | $0.10 | 10 |
| Parametric probesearch off | answer | $0.005 | 200 |
| Site and search data | |||
| Crawl, HTML page | page | $0.0002 | 5,000 |
| Crawl, JS-rendered page | page | $0.002 | 500 |
| SERP / keyword / backlink lookup | ≤100 rows | $0.005 | 200 |
| Agent squads | |||
| Why diagnosis | diagnosis | $0.05 | 20 |
| Answer analysis: sentiment, framing, claimsagent squad | answer | $0.002 | 500 |
| Prompt research: buyer questionsagent squad | call | $0.02 | 50 |
| Always free | |||
Reports, stored answers, jobs, quotes (dry_run), statistics, account, bot log reports | call | $0 | ∞ |
Everything on a test key (sk_test_…) | call | $0 | ∞ |
Consumer-app engines: chatgpt_ui, google_ai_mode. Native API + web search ($0.05): chatgpt, gemini, perplexity, deepseek, qwen, kimi. Premium ($0.10): claude, grok. How holds and refunds work: Billing.
Pricing questions
What does “credits are dollars” mean?
Your balance and every price are in US dollars. There's no made-up credit unit to convert. Every API response includes what the call cost (charged_usd) and your remaining balance.
Do I need a card for the free plan?
No. Sign in with Google, GitHub or email and you get $1 of usage every month with every engine and every feature, and one project. Add a card only when you top up or pick a plan.
What counts as an answer?
One response from one engine to one prompt. A panel of 10 prompts × 3 engines × 5 samples is 150 answers. Reports, stored answers, quotes and statistics are free to read as often as you like.
Am I charged for failures?
No. A run holds its quoted price up front, then charges only the answers that succeeded. Provider errors, timeouts and engines that aren't configured are refunded automatically.
Queued or live?
Consumer-app engines (ChatGPT app, Google AI Mode) can run queued: results arrive in minutes instead of seconds, for $0.003 instead of $0.01 per answer. Scheduled tracking panels should almost always run queued.
Do plans lock features?
No. Every plan, including Free, gets every engine and every MCP, CLI and API feature. Plans buy usage at a bonus rate plus projects, seats and agency features. Unused plan usage rolls over for one month; pay-as-you-go top-ups never expire.
What are test keys?
Keys starting sk_test_ run against mock engines that return realistic, varied answers with the same shapes and errors as live keys. They're always free, so you can build and test your agent integration without spending anything.
Can I use my own provider keys?
On Agency and Enterprise you can bring your own OpenAI, Anthropic, Google or Perplexity keys and pay only a platform fee of 20% of list price for native-API engines. The CLI's --local mode also runs panels on your machine with your own keys.
Start with $1 on us.
Every engine, every feature, one project. No card needed.