Limits
Budgets that bound cost, not a plan.
Every figure below is a real ceiling your AI can hit. None of them are a plan tier — there isn't one to upgrade out of.
Why limits exist #
These budgets bound what the database costs to serve. They are not a plan tier, and there is no paid tier that lifts them — the CRM is free. A single workspace can hold millions of company rows, and an unbounded agent loop or a runaway scrape would cost the same whether or not anyone paid for it, so the limits stayed after billing was removed.
If a call hits one, see Troubleshooting for what the specific message means and what happens next.
Daily budgets #
These are the budgets an agent most often hits. Every tool call draws against the tool-call budget; company detail reads and list rows also draw against their own tighter, data-volume budgets on top of that. A few narrower actions — report gaps, probe-audit writes, document-file reads — carry their own tighter budgets not shown in the table below. Hit any of them and your call is refused, not queued.
| Action | Per minute | Per day |
|---|---|---|
| Tool calls | 300 | 2,000 |
| Company detail reads | 120 | 1,000 |
| List rows (row-equivalents) | 2,500 | 10,000 |
| Report saves | 10 | 200 |
| Record writes (create / update) | 10 | 200 |
These are rolling windows, not a fixed reset time — a “per day” budget frees up gradually as calls from roughly 24 hours ago age out, rather than resetting all at once at UTC midnight. In practice this means an agent that hits a limit usually resumes on its own within the window, with no action needed from you.
The monthly load cap #
Loading companies from the shared national universe into your workspace is metered separately from the budgets above: 5,000 rows per calendar month, resetting at the start of the next month (UTC). This is the cost of pulling from a dataset with tens of millions of rows into your own tenant — it has nothing to do with what you've already loaded.
Companies already on your workspace are never affected by this cap — it only limits how many new rows you can pull in. New accounts also get a one-time free sample load that draws from its own separate allowance, so trying SoloSearcher for the first time doesn't compete with your monthly budget.
The exact message when the cap is hit
“Monthly load cap reached (5000/5000 rows loaded this month).” The first number is your current total for the month, so it always matches the second when this message appears.
Same framing as everything above: this bounds database cost, it is not a paywall, and there is no plan that raises it.
Field caps #
Every text field an AI can write on a company, contact, or listing record has a character cap, to keep a runaway model from filling a record with unbounded text in one call. A saved report's markdown body is deliberately exempt from these caps — it gets its own much larger 256KB limit instead, since a real analysis can run long:
- Names and titles — company, contact, and listing names — up to 500 characters.
- Short text — phone, city, source, external ids — up to 300 characters.
- Prose — notes, process, buyer history, fit reason, key risk — up to 4,000 characters.
- URLs — up to 2,000 characters, deliberately longer than short text since query strings run long.
Hit one of these and the write is rejected outright rather than truncated — your AI sees the cap and can retry with a shorter value.