Providers
Language Models
This section is only available to organization admins. It lists the large language models (LLM) and embedding models used by the organization.
Organization admins can:
- Create new LLM and embedding models — see Admin Panel - LLM Provider for instructions on creating LLM and embedding providers.
- Activate or deactivate existing LLM and embedding models for the organization
- Set a default LLM for the organization
Embedders
Embedding models appear below the LLM section.
Import and Export
You can import or export the list of LLM and embedding models of your organization in JSON format by clicking the following buttons. The export is generated as a zip file and also contains the logos and cost metadata of the models and embedders.
To import or export, click on the button at the top right of the LLM or embedders section
Website Providers
Website providers power the Web Search tool in the LLM Chat. Configure one here to let the LLM search the web during a conversation. Only one provider is active at a time — the one marked as default.
Adding a provider
Scroll to the Web Search section, select a provider, and follow the three steps shown in the card:
- Follow the link to the provider's dashboard and obtain an API key.
- Paste the key into the field and click Test & Save. A live query validates the key before it is stored.
- The provider is now active and set as default.
Switching providers
You can save credentials for multiple providers. Use the Set as Default button to switch the active one at any time. Credentials for inactive providers are kept and can be reactivated without re-entering the key.
Available providers
Linkup — EU-hosted, GDPR-compliant, zero data retention. Recommended. app.linkup.so
Brave Search — Independent search index with automatic language detection. api.search.brave.com
Serper — Google Search results via API. serper.dev
Exa — Neural/semantic search with highlight extraction. exa.ai/dashboard
Image Generation
Configure the provider used by the image generation tool. Once configured, agents and the LLM Chat can generate images from text prompts. Only one provider is active at a time — the one marked as default.
Adding a provider
Scroll to the Image Generation section, select a provider, and follow the three steps shown in the card:
- Follow the link to the provider's dashboard and obtain an API key.
- Paste the key into the field, select a model, and click Test & Save. A live query validates the key before it is stored.
- The provider is now active and set as default.
Switching providers
You can save credentials for multiple providers. Use the Set as Default button to switch the active one at any time. Credentials for inactive providers are kept and can be reactivated without re-entering the key.
Available providers
Black Forest Labs (Flux) — EU-hosted, GDPR-compliant. Generates images with Flux models. Recommended. auth.bfl.ai
OpenAI — Generates images with DALL-E 3 or GPT Image 1 via the OpenAI API. platform.openai.com
Azure OpenAI — Generates images using DALL-E deployed on Microsoft Azure OpenAI.
Google Vertex AI — Generates images with Google Imagen via Vertex AI.
Speech to Text
Speech to Text (S2T) providers power the microphone button in the chat interface, allowing users to dictate messages instead of typing.
Configure a provider
To enable speech-to-text, expand a provider card, fill in the required credentials, and click Test & Save.
- Click a provider card to expand it.
- Enter the API key (or URL/token for On Premise).
- Click Test & Save — a "Connection successful" message confirms the provider is ready.
Switch the default provider
Click Set as default next to any configured provider. The selected provider is then used for all microphone transcriptions in the organisation.
Available providers
On Premise — Self-hosted transcription service — no data leaves your infrastructure. Supports speaker diarization.- OpenAI — Transcribes audio with OpenAI Whisper via the OpenAI API. platform.openai.com
Mistral — Transcribes audio with Mistral Voxtral via the Mistral API. EU-hosted. console.mistral.ai