Plans and limits
What each Fond plan allows: websites, tracked queries, runs, assistants and models, monthly actions, integrations and rank tracking.
Every number below is a limit that the app enforces, generated from the product's own configuration.
| Limit | Audit | Growth | Premium |
|---|---|---|---|
| Websites | 1 | 1 | 1 |
| Tracked queries per website | 10 | 50 | 100 |
| Runs | 1 run total | Up to 7 runs per week | Up to 7 runs per week |
| Runs on a schedule | No, not on a repeating schedule | Yes, on a schedule Fond sets | Yes, on a schedule Fond sets |
| Assistants | OpenAI | Claude, Gemini, OpenAI, Google AI Mode | Claude, Gemini, OpenAI, Google AI Mode |
| Assistants per run | 1 | 4 | 4 |
| Models your plan allows | 1 | 3 | 7 |
| 0 Reddit posts per month | 5 Reddit posts per month | 15 Reddit posts per month | |
| Citation reports per month | 0 | 3 | 6 |
| Schema optimizations per month | 0 | 2 | 4 |
| Google Analytics connections | 0 | 1 | 1 |
| Cloudflare edge connections | 0 | 0 | 1 |
| Rank trackers per website | 1 | 3 | 10 |
| Keywords | 50 tracked keywords | 250 tracked keywords | 1000 tracked keywords |
| Search results read per check | 10 | 20 | 20 |
| Devices per keyword | 1 | 1 | 2 |
| Fastest rank-check schedule | Paid plans only | Weekly | Daily |
A monthly limit follows your billing cycle, not the calendar month. The three monthly limits above add up to:
- Audit: 0 monthly actions
- Growth: 10 monthly actions
- Premium: 25 monthly actions
How to read a few of these
Plan & usage
The Plan & usage page in the app meters the limits that you can use up, such as websites, tracked queries and runs. It lists some ceilings with nothing counted against them, such as which assistants your plan allows. It leaves out some rows, among them Assistants per run and Models your plan allows.
The Plan & usage page shows no price. The plan page in Fond shows what each plan costs. An organization billed through a marketplace sees its price there.
Runs
The Runs row is a ceiling, not how often runs happen. Fond runs a scheduled website every other day, unless it has set another interval. Runs that you start count against the same weekly ceiling.
The audit plan has one run, which Fond starts when you finish the website setup. If a failure in that run came from a source that Fond no longer collects from, Fond excuses one replacement run, once for each website. A question and assistant that failed in both of the ways that Fond measured it gets no replacement. The REST API reference tells you when the excusal applies.
Assistants and models
Assistants names the assistants that a run can use. Assistants per run is how many one run can use. Models your plan allows counts the models across those assistants. It is a count, because model names change with no change to your plan. It does not promise that you choose each model.
Some assistants have no model to choose, so they add nothing to the count. Fond can also change how it measures an assistant, for everyone or for one website. Then a model that your plan allows can have no effect. Your website's runner configuration shows a model list where there is a choice, and an on/off switch where there is not.
Rank tracking
The four rank tracking limits are separate because they multiply. A keyword checked daily, on two devices, reading 100 results costs six hundred times one checked monthly, on one device, reading 10. The keyword total is the main number.
Still rolling out
Edge measurement is still rolling out. It covers the Cloudflare edge connections row above. This row gives what your plan allows once Fond switches it on for your account. Until then, it is not in your dashboard or on the Plan & usage page.
Search and rank tracking is still rolling out. It covers the Rank trackers per website, Keywords, Search results read per check, Devices per keyword and Fastest rank-check schedule rows above. These rows give what your plan allows once Fond switches it on for your account. Until then, it is not in your dashboard or on the Plan & usage page.
Custom plans
Each custom plan is a separate agreement with its own figures, not the ones on this page. Fond can set every limit above for each organization, except the Runs on a schedule row. The plan that the agreement starts from sets that row.
What this page does not cover
- Prices and your own usage, which are not limits. The Plan & usage section above says where each one is.
- Internal spending ceilings on what Fond's own agents spend for you at third-party data providers. They limit nothing that you do by hand, and Fond does not publish them.
- What the limits are for. How Fond works explains what a run is, and why the number of assistants changes what a rate means.
Share of voice
How often AI answers mention your brand in each group of queries, by query tier or by your topic clusters. A query counts once if any answer to it mentioned you.
REST API Reference
Reference for the Fond REST API: authenticate with an API key, retrieve brand-visibility analytics, and trigger query runs under /api/v1/.