Reading Job charges
Thecharge_credits field on a Job reports the credits charged so far, after refunds, when you read the Job. Credits held but not yet charged are not counted, and later charges or refunds can change the amount after the Job finishes. A charge that is unknown is null, not zero. Jobs from the original routes do not include the field.
API products
These are the default prices. Single-record, batch and uploaded-file Intelligence use the same tier price per completed row. Choose the tier on each API request. The People API supportsmicro, low, medium, high and xhigh.
People Intelligence
One
xhigh unit is the starting research budget for one person: 250 credits. Larger max_credits budgets buy more units. See Atlas Investigations.
Company Intelligence
People research and the legacy full organization chart option cost nothing extra. Bulk Intelligence costs the same per row as People and Company Intelligence.
Contact lookups
Each found email, phone number or reverse match counts as one result. A lookup with no result is free. An email counts as found unless verification marks it invalid, so risky, catch-all and unknown emails are charged. Selecting more than one contact type charges for each type found.
Search
People and company Search Jobs charge for the results they retrieve while building their saved results. Reading the Job or its items does not charge again. Exporting Job results and saving a list of IDs are free.
The Search Export row applies to legacy exports. Legacy search billing explains when older searches charge for reading and exporting results.
Verification
Email verification does not charge for an
unknown verdict. Age Verification uses the high-tier research price. Workflow verification blocks have separate prices below.
AI agents
These tools use workflow blocks or the original API routes.
LinkedIn Enrichment also charges when it finds no profile.
Atlas Investigations
Atlas research investigations are priced in units. Org chart tasks use the per-person pricing below. You choose a run’s budget withmax_credits, and more credits buy a deeper investigation. A run costs the same whether it comes from the Atlas API or from People Intelligence.
In the Atlas API, set
engine to "scout" or "xhigh". People Intelligence supports tier: "xhigh".
max_creditsdefaults to one unit.- A run starts only when your balance covers its
max_credits. It never costs more than that, and credits it doesn’t use go back to your balance. - A run that stops early is charged for the research it did, and at least one unit.
- A run that reaches its
max_minuteslimit is charged for the research it did, rounded up to the next 50 credits. This can be less than one unit. - A run the agents decline to research is free. See Declined runs.
Org charts
Investigations withtask: "org_chart" cost 20 credits per unique person placed, by default. Teams do not count. Updating an already-paid person or continuing the same chart does not charge for that person again; newly placed people are charged.
Org chart tasks have no per-run credit cap and do not accept max_credits. Their charge grows with the number of people placed, rather than research units. The research-budget limits above apply to ordinary investigations.
Billed tools
Some tools the agent uses while it researches cost credits. They are paid for out of the run’s research budget (max_credits), not charged on top of it, so a run still never costs more than its budget. Other runs don’t pay for them, but follow the team’s settings below.
The Browser works through web pages that need clicks or typing, such as a public records search form. Transcription turns the audio of a podcast episode, a video or a recording the agent downloaded into text that says who spoke and when. A recording is charged once, however much of it the agent reads.
- You’re charged for a call only when it succeeds: a Browser task when it finds what it was asked for, and a transcription when it returns text. A Browser task that finds no match isn’t charged, nor is audio with no speech in it, and captions are free.
- Team owners and organization admins turn each tool on or off, and can cap how much a run uses, under Paid sources in Settings. Every tool is on by default.
Monitors
Monitors are enabled per organization. Contact support to turn them on.
A probe checks each active row that already has a baseline. The first check fills missing watched values; rows that supply every watched value skip it. A re-check researches a row after its probe finds a likely change.
The tier price is the People Intelligence price for
lead monitors and the Company Intelligence price for company monitors. A probe, first check or re-check charges only when it finishes. A LinkedIn check charges only when it finds a change, regardless of how many fields change.
Lattice
Lattice is in beta and available to Enterprise customers, on request. Talk to sales to try it.
Only rows added to the graph incur the listed charge.
Workflow blocks
Blocks you add to a workflow charge per row or per call as listed.Data sources
Each subdomain counts as one result. A Google Maps page contains up to 20 results. TikTok Get Following fetches every page.
Web scraping
Premium proxies and JavaScript rendering add to the base price.
AI text generation
Transform data
A row whose output fields are all filled by rules pays only for its list matches.
Verification blocks
Email verification has a minimum charge of 0.1 credit per run and does not charge for unknown verdicts. Phone verification skips rows with a blank phone number or no name and does not charge for those rows.
When credits are charged
Successful lookups and workflow steps charge as they complete. Search charges for the results it retrieves. Completed steps stay charged if a later step fails.What’s free
API
Workflows
Web app
Getting the Most from Your Credits
A few habits help every credit go further.- Filter before you research. Remove duplicates and rows you don’t need first, so you pay only for records you’ll use.
- Test on a small batch. Run a new workflow on 10 to 20 records before you run it on your full list.
- Set a budget for Atlas. Start with one unit with
max_credits, and continue the investigation if you need more. Anything a run doesn’t spend comes back to you. - Give monitors what you know. Include the values you already have when you add rows, so they skip the paid first check. Switch off rows you no longer need to watch.
- Ask for the results you’ll use. Set
max_resultsto what you need. Search Jobs charge for the results they retrieve, not when you read the saved items. - Keep what you’ve paid for. Downloading results and reading Atlas investigations, reports and exports are free, so save your results instead of running the same lookup again.
- Cap billed tools. If you’re a team owner or admin, you can cap how much a run spends on billed tools under Paid sources in Settings.
- Keep an eye on spending. Usage & API Logs shows the credits you use by endpoint and over time.