Pricing
You pay for what you ingest and what you ask.
The people using your software never need a Revo account, so we count what the API does rather than how many people touch it. Every call returns its own cost under usage, so you can price a feature or cap a customer without instrumenting anything.
Free
$0
no card, no commitment
For a side project, or for evaluating the API against your own data before you write an integration.
- 10,000 ingest requests / month
- 1,000 questions / month
- 5,000 resolved entities
- 1 project
- Unlimited end users
- Community support
Pay as you goPopular
Credits
prepaid, with automatic top up
For production traffic. You buy credits up front and calls draw them down, with auto top up so a busy week never stops your product mid-request. No ceilings on anything.
- Prepaid credits, drawn down per call
- Automatic top up at a threshold you set
- Unlimited ingest requests
- Unlimited questions
- Unlimited resolved entities
- Unlimited projects
- Unlimited end users
- Usage analytics and spend alerts
- Private Slack support
Enterprise
Custom
annual agreement, volume rates
Same unlimited engine, priced for committed volume, with the operational and contractual terms a platform team asks for before putting your memory in front of their own customers.
- Everything in Pay as you go
- Volume pricing on an annual commitment
- Dedicated or in region deployment
- SLA and support terms
- DPA, retention policy and legal hold
- Audit logging
- Custom ontology for your domain
- SSO
Compare plans
Every limit, side by side.
| Feature | Free | Pay as you go | Enterprise |
|---|---|---|---|
| Usage and limits | |||
| Price | Free | Metered on prepaid credits | Custom |
| Ingest requests / month | 10,000 | Unlimited | Unlimited |
| Questions / month | 1,000 | Unlimited | Unlimited |
| Resolved entities | 5,000 | Unlimited | Unlimited |
| Projects | 1 | Unlimited | Unlimited |
| End users | Unlimited | Unlimited | Unlimited |
| History retention | 30 days | Unlimited | Contractual |
| Billing | None | Prepaid credits, auto top up | Invoiced on an annual commitment |
| Engine | |||
| All four endpoints | Yes | Yes | Yes |
| Answers with evidence | Yes | Yes | Yes |
| Per call cost in the response | Yes | Yes | Yes |
| Entity resolution across sources | Yes | Yes | Yes |
| Temporal history and conflict retention | Yes | Yes | Yes |
| Permissioned retrieval | Yes | Yes | Yes |
| Custom ontology for your domain | Not included | Not included | Yes |
| Support and enterprise | |||
| Support | Community | Private Slack | Private Slack and SLA |
| Usage analytics | Not included | Yes | Yes |
| Dedicated or in region deployment | Not included | Not included | Yes |
| Audit logging | Not included | Not included | Yes |
| DPA, retention policy and legal hold | Not included | Not included | Yes |
| SSO | Not included | Not included | Yes |
*Tokens are issued by hand while the API is in early access, on every tier including the free one. There is no dashboard and no self serve signup yet, so every plan starts with a conversation. Self serve is dated on the roadmap.
What counts
Three meters, defined precisely.
You should be able to work out the bill before writing any code. These are the three things we count.
Ingest requests
One data point is one call to POST /ingest: an email, a transcript, a ticket, a note.
Ingestion is idempotent by content hash, so re-sending unchanged content costs nothing. A crashed backfill can resume by re-running the same loop.
Questions
One call to /recall or /answer.
We read a roughly fixed amount per question whether the record holds a hundred thousand tokens or ten million, so the bill does not grow with the history.
Resolved entities
The people, companies and initiatives the graph currently maintains inside one project.
Metered on what the graph maintains rather than on bytes stored.
Why it stays flat
The bill does not grow with the history.
The usual alternative is to send the whole history to the model on every question, so cost climbs with every message you have ever stored. We read a roughly fixed amount per question whether the record holds a hundred thousand tokens or ten million.
23x
cheaper per question than sending 1M tokens of history, on the same model.
232x
cheaper at 10M tokens, where no production model can hold the context at all.
~96%
less context read per question at 1M tokens of history.
Expressed as multiples rather than per question prices, because what we charge is not what it costs us to run.
Traffic that does not map onto a tier?
Most early access integrations are usage based agreements rather than a plan off this page. Tell us the volumes and the shape of your integration and we will model the bill with you.
FAQ
What a bill is made of, and what early access actually means for getting started.