Integration Intelligence
AI Integrations
Verified patterns for connecting AI models and APIs to apps, data, and workflows — from official product and developer documentation only.
24 integration profiles
Agent & tool calling
-
Amazon Bedrock Agents
Amazon Web Services
AWS documents Bedrock Agents as an orchestration layer that plans multi-step tasks, optionally uses Knowledge Bases, and invokes Action Groups (APIs/Lambda) to integrate Bedrock models with enterprise backends.
-
Anthropic Tool Use (Messages API)
Anthropic
Anthropic documents tool use on the Messages API so Claude can request structured tool invocations that the client executes and returns, forming the integration contract between Claude and external systems.
-
Gemini Function Calling
Google
Google AI documents function calling for Gemini so applications can declare functions the model may call, then execute those functions and return results—Google’s primary integration pattern for connecting Gemini to app logic and APIs.
-
OpenAI Custom Actions (GPTs)
OpenAI
OpenAI documents Custom Actions for GPTs so a GPT can call external HTTP APIs via an OpenAPI schema, enabling tool-style integrations from ChatGPT into third-party systems when configured by the builder.
-
Vertex AI Extensions
Google Cloud
Google Cloud documents Vertex AI Extensions as a way to connect generative models on Vertex AI to external APIs and enterprise systems for tool-augmented workflows under GCP IAM.
Automation platforms
-
Make ↔ OpenAI Integration
Make
Make (formerly Integromat) documents OpenAI modules so scenarios can call OpenAI APIs as steps in visual automations connecting hundreds of other apps.
-
n8n ↔ OpenAI Integration
n8n
n8n documents OpenAI nodes so workflow automations can call OpenAI models as steps alongside other app integrations in self-hosted or n8n Cloud environments.
-
Zapier ↔ OpenAI Integration
Zapier
Zapier documents an OpenAI app/integration that lets Zaps send prompts and receive completions from OpenAI models inside multi-app automation workflows, using OpenAI credentials configured in Zapier.
Developer tools
-
GitHub Copilot IDE Integration
GitHub
GitHub documents Copilot as an AI pair-programmer integrated into supported IDEs and editors (VS Code, JetBrains, Visual Studio, Neovim, and related surfaces) plus GitHub.com chat experiences where enabled.
Enterprise SaaS
-
Azure OpenAI with Microsoft Power Platform
Microsoft
Microsoft documents integrating Azure OpenAI Service with Power Platform (Power Automate, Power Apps, Copilot Studio patterns) so organizations can call Azure-hosted models from low-code workflows under Azure identity and governance.
-
Gemini for Google Workspace
Google
Google documents Gemini for Google Workspace as generative AI features integrated into Gmail, Docs, Sheets, Meet, and related Workspace apps for eligible Workspace editions.
-
Microsoft 365 Copilot
Microsoft
Microsoft documents Microsoft 365 Copilot as an AI assistant integrated into Microsoft 365 apps (Word, Excel, PowerPoint, Outlook, Teams, and related surfaces) grounded on Microsoft Graph data with enterprise controls.
-
Salesforce Agentforce / Einstein AI
Salesforce
Salesforce documents Agentforce and Einstein generative AI capabilities that integrate AI agents and assistants into Salesforce CRM clouds with Salesforce Trust and admin controls.
-
Slack AI
Salesforce / Slack
Slack documents Slack AI features (summaries, search, and related assistants) that integrate generative AI into Slack workspaces under Slack’s product and admin controls.
Orchestration frameworks
-
LangChain ↔ Anthropic Integration
LangChain
LangChain documents Anthropic chat model integrations so Claude can be invoked through LangChain abstractions using Anthropic API credentials and Messages API semantics.
-
LangChain ↔ OpenAI Integration
LangChain
LangChain documents first-class OpenAI chat and embeddings integrations so application code can call OpenAI models through LangChain’s runnable/chain abstractions while using OpenAI API credentials.
-
LlamaIndex ↔ OpenAI Integration
LlamaIndex
LlamaIndex documents OpenAI LLM and embedding integrations used to power indexing and query pipelines (RAG) against OpenAI models with OpenAI API keys.
-
Vercel AI SDK Multi-Provider Integration
Vercel
The Vercel AI SDK documents a unified TypeScript interface for streaming UI and server helpers across multiple model providers (OpenAI, Anthropic, Google, and others as listed in its provider docs).
RAG & data platforms
-
Amazon Bedrock Knowledge Bases
Amazon Web Services
AWS documents Bedrock Knowledge Bases as a managed RAG integration that connects foundation models on Bedrock to enterprise data sources for retrieval-augmented generation under AWS IAM and regional controls.
-
Cohere RAG / Chat with Connectors
Cohere
Cohere documents retrieval-augmented generation patterns (including chat with documents/connectors where offered) that integrate Command models with retrieved enterprise or web knowledge under Cohere’s API.
-
Snowflake Cortex AI
Snowflake
Snowflake documents Cortex AI as SQL and platform functions that bring LLM and ML capabilities into the Snowflake Data Cloud, integrating model inference with governed warehouse data.
Model serving
-
Databricks Mosaic AI Model Serving
Databricks
Databricks documents Mosaic AI Model Serving as a way to deploy and query ML/LLM endpoints inside the Databricks Lakehouse, integrating inference with Unity Catalog governance and Databricks workflows.
-
Hugging Face Inference Endpoints
Hugging Face
Hugging Face documents Inference Endpoints as a managed deployment integration for Hub models, exposing HTTPS inference endpoints that applications call with HF tokens.
-
Replicate Model Deployments
Replicate
Replicate documents running and deploying machine learning models via versioned containers with an HTTP API, integrating third-party apps to model versions hosted on Replicate.
APIs & SDKs → · AI Models → · Compliance → · Technologies → · Compare →