1
Raw data
Retrieving the source itself — loading pages, expanding discussion threads, pulling transcripts. Richer sources take more retrieval work: a discussion thread with hundreds of replies loads many pages, a simple article loads one.
Pricing details
What things cost
Every task your agent runs is metered in credits — the same credits across all plans and top-up packages. This page shows what typical actions cost and why.
2
AI processing
Every source is cleaned, indexed line by line, and scored for relevance to your query, so your agent can cite exact passages instead of re-reading everything. Media analysis (video transcription, image description) runs only where it applies.
3
Metered per task
You pay for work actually done — no seats, no minimums. Re-reading data you already collected is free, and re-fetching a recently ingested URL costs a fraction of the first fetch thanks to caching.
Figures are typical averages from live usage — actual billing is metered per operation, so a small thread costs less and a huge one costs more. Euro examples use the Medium top-up rate (€0.0002 per credit); larger packages bring the per-credit price down.
Search
One search query, including relevance scoring of the results.
| Action | What's included | Credits | ≈ per request |
|---|---|---|---|
| One web search query, results scored for relevance | ~12 | €0.0025 | |
| YouTube | One video search query | ~13 | €0.0026 |
| Google Maps | One places query, geo-targeted automatically | ~14 | €0.0028 |
| TikTok | One short-video search query | ~20 | €0.0040 |
| Threads | One social search query | ~20 | €0.0040 |
| One community search query | ~28 | €0.0056 |
Reading sources
Fetching a URL: raw retrieval plus full AI processing — cleanup, indexing, relevance.
| Action | What's included | Credits | ≈ per request |
|---|---|---|---|
| PDF document | Text extraction and indexing | 3 | €0.0006 |
| Web page or article | Full scrape, cleanup, line-by-line index | ~17 | €0.0033 |
| X (Twitter) post | Post with author, engagement and replies | ~21 | €0.0042 |
| Facebook post | Post with author, engagement and comments | ~21 | €0.0042 |
| First and foremost person profiles — plus company pages and posts | ~21 | €0.0042 | |
| Threads post | Post with author, engagement and replies | ~21 | €0.0042 |
| Bluesky post | Post with author, engagement and replies | ~21 | €0.0042 |
| Pinterest pin | Pin with description and engagement | ~21 | €0.0042 |
| Twitch content | Channel or clip details | ~21 | €0.0042 |
| Snapchat content | Public profile or story details | ~21 | €0.0042 |
| Truth Social post | Post with author, engagement and replies | ~21 | €0.0042 |
| YouTube video | Timestamped transcript plus video details | ~42 | €0.0084 |
| Instagram post | Post with replies; images described by AI where present | ~55 | €0.0110 |
| TikTok post | Post with replies and video transcription | ~59 | €0.0118 |
| Reddit thread | Thread with replies — billed for the reply pages actually loaded | ~120–200 | €0.0320 |
Answers
Working over data you already collected: cited answers grounded in your sources.
| Action | What's included | Credits | ≈ per request |
|---|---|---|---|
| Ask — per source | Each data record consulted while answering your question | ~4 | €0.0009 |
| Ask — typical question | A question answered over ~30 collected sources, with citations | ~150 | €0.0310 |
Why do sources cost different amounts?
Because they take different amounts of work. A PDF is parsed locally; a social post is a couple of retrieval calls; a long discussion thread expands into dozens of pages of replies. Billing follows the actual work, so simple things stay cheap.
What about repeated work?
Data you've collected is stored and re-readable for free. Re-fetching the same URL soon after ingestion serves from cache and only pays for a light relevance pass — typically 1–2 credits.
Are credits the same everywhere?
Yes — a credit buys the same work whether it came from a monthly plan or a top-up package. Only the price per credit differs: larger packages and plans are cheaper per credit.