Pricing
Start for free. Upgrade to get the capacity that exactly matches your team's needs.
Find out how Vellum helps you build powerful agents.
Designed for anyone starting to build agents and test ideas.
Designed for teams scaling agent deployments across the org.
Built for large orgs needing flexibility, scale, and governance.
Frequently asked questions
Agent builder credits represent how many messages you can send while creating or testing agents in Vellum. Each message in the builder counts as 1 interaction, and your plan defines how many you get per month.
The context storage is powered by a built-in vector database that your agents can use for retrieval (RAG). This lets your agents access relevant knowledge during runs. Higher tiers allow you to store and search more docs for richer, more accurate responses.
Hosted agent apps are live, shareable links automatically generated for your agents. Each one gives your agent its own hosted interface; no setup or code required. You can share the link with your team.
Parallel agent runs determine how many people can use your hosted agent link at the same time. Higher plans allow more users to interact with your agent concurrently without waiting for others to finish.
Enterprise gives you everything in Business plus features for governance and scale: RBAC, isolated environments (dev/staging/prod), Slack support & SLAs, Prompt Management, Evals, and the option for VPC installation and custom legal terms.
The Agent Console lets you debug and test your agents as you build. You can see detailed traces of each run, view inputs and outputs step-by-step, and replay executions to understand how your agent made decisions.
The data retention period defines how long your logs and execution data remain viewable in Vellum. When the period ends, older data will be automatically hidden.
When you hit your storage cap, your newest agent runs still work, but older execution data will automatically be hidden once your retention period ends. Upgrading increases both storage size and how long you can access your data.
Yes. You can review execution history, see performance metrics, and use traces in the console to spot and fix issues quickly, even after your agents are live.
Yes. With isolated environments, you can safely test agents in development or staging before promoting them to production, keeping your workspace organized and reducing risk. This feature is available only on enterprise plans.