Integrations

Connect your
AI Agents
to the tools your company already uses

Integrate systems, data, and applications so your Secure AI Agents can query information and execute actions within the limits defined by your company.

33 integrations in AI

Google Gemini

Use Google's Gemini models to generate and analyze content.

Microsoft Copilot

Connect Microsoft Copilot capabilities to the agents' flows.

Amazon Bedrock

Use AI models available on Amazon Bedrock.

Google Vertex AI

Use Vertex AI models and capabilities.

Azure OpenAI

Use OpenAI models hosted on Azure.

Pinecone

Store and search vectors for semantic search in Pinecone.

Weaviate

Store and search vectors in the Weaviate database.

Qdrant

Store and search vectors in the Qdrant database.

Milvus

Store and search vectors in the Milvus database.

Chroma

Store and search embeddings in Chroma.

OpenAI

Use OpenAI models to generate text, images, and embeddings.

Anthropic

Use Claude models through the Anthropic API.

Claude

Connect Claude to your agents' flows.

Google AI Studio

Use Google models through Google AI Studio.

Cohere

Generate text, embeddings, and reranking with Cohere.

Mistral AI

Use Mistral AI models to generate and analyze content.

Groq

Run AI models with low latency on Groq.

Together AI

Run open AI models on Together AI.

Replicate

Run AI models hosted on Replicate.

Hugging Face

Use Hugging Face models and datasets.

Stability AI

Generate and edit images with Stability AI models.

ElevenLabs

Generate realistic voices and audio with ElevenLabs.

Deepgram

Transcribe audio and calls with Deepgram.

AssemblyAI

Transcribe and analyze audio with AssemblyAI.

Speechmatics

Transcribe audio in many languages with Speechmatics.

OpenRouter

Access many AI models through a single API with OpenRouter.

LangChain

Connect chains and tools built with LangChain.

LangSmith

Track and evaluate LLM runs in LangSmith.

Langfuse

Monitor and evaluate LLM applications with Langfuse.

LlamaIndex

Connect indexes and prepared data with LlamaIndex.

CrewAI

Connect agent crews built with CrewAI.

AutoGen

Connect agents built with AutoGen.

Pinecone Assistant

Use Pinecone knowledge retrieval assistants.

Control and security

Connecting does not mean granting everything.

Each integration exposes specific capabilities. What a Secure AI Agent can query or execute depends on the permissions, policies, and autonomy level defined by your company.

Integration: Defines which systems and information sources are available to the agent.
Permissions: Determine which data and actions can be accessed within each connected system.
Autonomy: Defines how far the agent can act alone and when human approval is required.

The availability of integrations, features, and actions may vary according to the environment configuration and the permissions granted.