AI Provider
Connect your organization's AI provider so every Encatch AI feature runs on your account and billing
Connect your organization's own AI provider account, and every Encatch AI feature runs on it instead of Encatch's built-in AI. Your provider bills you for the usage, and Encatch doesn't deduct any AI credits.
Why use your own provider
- Keep AI usage under your own agreement. Requests run under your organization's contract and data terms with the provider.
- Save your AI credits. While your provider is connected, Encatch AI features don't consume Encatch AI credits.
- Track AI spend in one place. Usage appears on your provider bill alongside the rest of your AI usage.
Which features use your provider
Once connected, all Encatch AI features run on your provider.
| Feature | What it uses AI for |
|---|---|
| Destination AI filter | Decides whether a response is forwarded to a destination |
| Destination AI filter (Jev engine) | The Jev engine for destination AI filters |
| Refine text | Cleans up a respondent's text answer in the SDK |
| Feedback Q&A | Answers questions about a form inside the SDK |
| Topic categorisation | Tags text answers with topics and aspects |
| Translation | Translates form text into other languages |
| Sample data | Writes realistic text for sample responses |
| AI form builder | Builds and edits forms from a chat |
| External Insights chat | Chat that analyses this project's feedback |
Jev engine
Destination AI filter (Jev engine) runs only on a Vercel AI Gateway connection; the default is not one, so choose Override.
How it works
- An organization admin enables Bring your own key under Manage Org & Projects → Settings → AI settings.
- In each project, add one or more provider connections and API keys under Settings → Global Settings → AI Provider.
- Encatch stores each key securely. After you save it, no one can view it in Encatch again, including admins.
- Set a default connection and model for the project, then map each AI capability to Use default, Override, or Disabled. You can add multiple connections and point different capabilities at different models or connections.
- When a capability runs, Encatch sends the request to the connection and model you configured. Results appear in Encatch as before; your provider bills you for the usage.
Bring-your-own-key is an organization-level choice. When it is selected, Encatch's built-in AI does not run for any project in that organization—projects must use your own provider connections for AI.
Connections and capability mapping are project-level. Projects under the same organization configure their own connections independently; you are not limited to one provider or one model across capabilities within a project.
If a project has no provider connections added (or no working configuration for a capability), AI features do not work in that project until you add connections and save your settings.
Availability
- Plans: Growth and Enterprise
- SDK changes: none needed
Next steps
- Read Data handling and security before you connect a provider.
- Follow Add an AI provider to set it up.
- See Which features use your provider for the full feature list.
Was this page helpful?