Baseten integration
Baseten is a ML model deployment platform that enables teams to deploy, scale, and monitor machine learning models in production. Build inference APIs for any model, auto-scale based on traffic, deploy custom models or popular open-source models, and integrate with your ML workflow. Perfect for deploying LLMs, computer vision models, recommendation systems, and custom ML pipelines.
Common Baseten use cases
- reduce infrastructure costs
- enable scale-to-zero
- remove unused models
- remove old deployments
- store HuggingFace token
- view deployment history
- audit configured secrets
- find available GPU types
- real-time ML predictions
- inventory deployed models
What you can do with Baseten
Each of these is one operation on one endpoint, available both as a workflow step and as a tool an AI agent can call on its own. Nothing bundles several operations into one action, so what a run did stays readable afterwards.
Most endpoints run on the free plan. A few, mainly the ones that publish on your behalf, need a paid plan.
Deployment, 12 operations
- POSTactivate_deployment
Activates a deactivated Baseten deployment to make it available for inference.
- POSTdeactivate_deployment
Deactivates a Baseten deployment, stopping all replicas and inference traffic.
- DELETEdelete_deployment
Deletes a specific deployment of a Baseten model by ID.
- GETget_deployment
Retrieves detailed information about a specific model deployment including status, replicas, and autoscaling settings.
- GETget_development_deployment
Retrieves the development deployment for a Baseten model.
- GETget_production_deployment
Retrieves the production deployment for a Baseten model.
- GETlist_deployments
Lists all deployments of a Baseten model with status, autoscaling settings, and configuration.
- POSTpromote_deployment
Promotes a deployment to production status, routing all inference traffic to that deployment.
- POSTpromote_development_deployment
Promotes the development deployment to production, creating a new production deployment.
- PATCHupdate_autoscaling_settings
Updates autoscaling configuration for a Baseten model deployment. Adjust replica counts and scaling behavior.
- POSTwake_development_deployment
Wakes a scaled-to-zero Baseten development deployment.
- POSTwake_production_deployment
Wakes a scaled-to-zero Baseten production deployment to avoid cold start latency.
Prediction, 6 operations
- DELETEcancel_async_request
Cancels a queued async prediction request on Baseten.
- GETget_async_request_status
Retrieves status metadata for an async Baseten prediction request.
- POSTrun_async_inference
Runs asynchronous inference on a Baseten model. Returns a request ID immediately for status tracking.
- POSTrun_development_async_inference
Runs asynchronous inference on a Baseten model's development deployment for testing.
- POSTrun_development_inference
Runs synchronous inference on a Baseten model's development deployment for testing.
- POSTrun_production_inference
Runs synchronous inference on a Baseten model's production deployment. Sends input and receives predictions in real-time.
Model, 3 operations
- DELETEdelete_model
Deletes a Baseten model and all deployments. Stops replicas and removes inference endpoints permanently.
- GETget_model
Retrieves detailed information about a specific Baseten model including deployment IDs and instance type.
- GETlist_models
Lists all deployed ML models in your Baseten workspace. Returns model IDs, names, deployment counts, and active deployment IDs.
Secret, 2 operations
- GETlist_secrets
Lists all secrets in your Baseten workspace. Returns names and creation dates, not values.
- POSTupsert_secret
Creates or updates a secret in your Baseten workspace for use in model deployments.
Instance type, 1 operation
- GETlist_instance_types
Lists all available GPU and CPU instance types for Baseten model deployments.
Use Baseten from Claude Code, Cursor or Codex
Connect Baseten in LiveContext, create an API key in Settings, then add the LiveContext MCP server to your client. Its Baseten operations become tools the client can call, next to every other integration you connected.
claude mcp add --transport http livecontext https://livecontext.ai/mcp --header "X-API-Key: YOUR_API_KEY"[mcp_servers.livecontext]
url = "https://livecontext.ai/mcp"
[mcp_servers.livecontext.http_headers]
"X-API-Key" = "YOUR_API_KEY"{
"mcpServers": {
"livecontext": {
"type": "http",
"url": "https://livecontext.ai/mcp",
"headers": {
"X-API-Key": "YOUR_API_KEY"
}
}
}
}How to connect Baseten
Create your workspace
Sign up free. Nothing to install.
Connect Baseten once
Add your Baseten connection in Settings. Every tool from this integration reuses it, and secrets are encrypted at rest and visible only to you.
Describe the job in chat
The workflow builds itself in front of you and picks the endpoints it needs. Drag the steps yourself on the canvas whenever you would rather be explicit.
Baseten integration FAQ
- What can you do with the Baseten integration?
- LiveContext exposes 24 operations on Baseten, covering deployment, prediction, model and secret. Common uses include reduce infrastructure costs, enable scale-to-zero and remove unused models.
Popular integrations
Instagram (Instagram Login)Instagram API with Instagram Login (graph.instagram.com). Manage a Business/Creator Instagram account…
TelegramMessaging platform API for building bots and integrations. Send messages, manage chats and channels, handle…
YouTube Data APIGoogle's video platform API for accessing video, channel, playlist, and comment data
TikTokSocial media platform API for content creators and businesses. Access user info, manage videos, get…
Twitter / XSocial media platform API for posting, reading, and managing Twitter/X content. Post tweets, read timelines,…
GmailGoogle email service API for sending, receiving, and managing emails. Access messages, threads, labels,…
xAIAI API from xAI (Elon Musk's AI company) powering Grok models. Access state-of-the-art language models with…
LinkedInProfessional networking platform API for managing profiles, connections, and content. Share posts, manage…