Supported providers
iBlueprint supports the following AI providers out of the box:OpenAI
GPT-4o, GPT-4, GPT-3.5, embeddings, DALL·E, Whisper, and more.
Anthropic
Claude 3.5 Sonnet, Claude 3 Opus, Haiku, and future Claude models.
Google Gemini
Gemini 1.5 Pro, Flash, and the full Gemini model family.
Ollama
Self-hosted open-source models via a local or remote Ollama instance.
Add an AI connection
1
Open AI Connections
Go to Settings → AI Connections and click Add connection.
2
Choose a provider
Select your AI provider from the dropdown. For Ollama, you will also need to supply a Base URL pointing to your Ollama instance (for example,
http://localhost:11434).3
Enter your API key
Paste your API key into the API key field. iBlueprint encrypts this value at rest and never exposes it in the UI after saving.
4
Fill in optional fields
- Base URL — override the provider’s default endpoint (useful for Azure OpenAI deployments or Ollama).
- Description — a short note so you can identify this connection later.
- Tags — free-form labels for grouping connections (for example,
production,team-alpha). - Expiry date — automatically deactivate the connection on a specific date.
5
Save and test
Click Save. iBlueprint will immediately test the connection and report whether the key is valid. You can re-run this test at any time from the connection detail page.
You can add multiple connections for the same provider — for example, separate keys for development and production, or keys scoped to different teams.
Share a connection
By default, a connection belongs only to you. You can share it with individual teammates or with an entire organization, and choose what they are allowed to do with it.
To share a connection:
- Open the connection from Settings → AI Connections.
- Click Share.
- Search for a user or select an organization.
- Choose a share type and an optional expiry date or maximum usage count.
- Click Confirm.
Track usage
Every time a Blueprint node calls an AI provider through a connection, iBlueprint records the event. From the connection detail page you can see:- Total API calls over time
- Tokens consumed per call
- Estimated cost per call
- Which Blueprints used the connection and when
- Any error responses returned by the provider
Manage Ollama (self-hosted)
Ollama connections work the same way as cloud providers, with two differences:- API key — Ollama does not require an API key by default. You can enter any non-empty string (for example,
ollama) to satisfy the field. - Base URL — you must supply the full URL to your Ollama server, including the port.
