Starter
Free for everyone
$0/ month
- 100K observability events+ $10/100K
- 24 CPU hours+ $0.35/hr
- 15 days of data retention
- Unlimited users, deployments, and projects
MASTRA STUDIO
Iterate Faster,
Ship Better Agents
Collaborate With Your Team
Gather expert feedback on agent performance and iterate on prompts with your team. Share traces so anyone can understand user behavior, debug issues, or assess response quality.
Start Building, Then Invite Your Team
Studio runs on localhost. When it's time, share it behind a URL — your team gets the same workspace.
Authentication
Lock down Studio with role-based access. Define who can view, edit, and execute.
Invite Domain Experts
Invite your team to traces, build datasets, run experiments, and refine prompts, even if they don't have access to the source code.Own your agent loop
Studio gives your team one place to learn from real usage, iterate faster, and improve every release.
Iterate Quickly
Test your agents in Studio. Switch models, tweak settings, and try different inputs.
Scorers
Score outputs with LLM-as-Judge or code. Run online evals in production or score any historical trace retroactively.
Datasets and Experiments
Define what good looks like, then measure against it. Build regression sets from production traces and run experiments after every change.Start building with a free account, or explore our pricing for more
Full PricingTeams
For growing teams
$250/ month
- 1M observability events+ $8/100K
- 250 CPU hours+ $0.25/hr
- 6 months of data retention
- Multiple teams, SSO, and SOC 2 docs
Enterprise
For teams at scale
Custom pricing
Custom volume and retention, with RBAC, audit logs, support and uptime SLAs, and a dedicated support engineer.
Mastra is powering the best AI teams
Case StudiesHow Replit Agent 3 creates thousands of Mastra agents every day
How SoftBank is restoring Japan's white-collar productivity using Mastra
How Sanity Built a Content Agent That Actually Understands Your CMS
How Marsh used Mastra to build Agentic Search for 100k employees
Frequently asked questions
What is Mastra Studio?
Mastra Studio is the workspace where your team observes, iterates and improves AI agents together. Mastra Studio monitors agent performance in production, tracking costs, latencies and quality scores. Follow execution traces to find exactly where an agent stalled or made the wrong call. Invite domain experts to review traces, build datasets, run experiments and refine prompts directly in the UI, even if they don't have access to the source code. Run Studio locally during development, then share it behind a URL so your team gets the same workspace.
How does Mastra Studio connect to my Mastra application?
Mastra Studio connects to your Mastra instance via a configurable server URL and runs at localhost:4111 during local development. Studio updates in real time as you edit your agents, workflows and tools. Share Studio behind a URL for team access when you are ready to deploy.
How does Mastra Studio help measure every agent run?
Mastra Studio gives you the data to understand what went wrong in production. Comprehensive metrics dashboards track costs, latencies and quality scores across every agent run. Traces and logs capture the full story behind every agent interaction, recording telemetry automatically so you can see what happened and how it all connects. Score outputs with LLM-as-Judge or code, and run online evals in production or score any historical trace retroactively.
How does Mastra Studio help improve agents over time?
Mastra Studio gives you the tools to find what changed between releases and fix it. Test agents directly in Studio by switching models, tweaking settings and trying different inputs. Edit and test prompts without writing code and see how changes affect outputs immediately, then serve different variants in production. Save agent configurations as versions, compare them side by side and roll back when something breaks. Browse MCP servers and swap configurations to find the right setup.
How do datasets and experiments work in Mastra Studio?
Mastra Studio datasets are collections of test cases for evaluating agents and workflows. Import items from CSV or JSON, define input and ground-truth schemas and pin to specific versions to reproduce experiments exactly. Run experiments with scorers to compare quality across prompts, models or code changes.
How does Mastra Studio support team collaboration?
Mastra Studio gives non-technical collaborators direct access to the same workspace developers use. Run Studio on localhost during development, then share it behind a URL so your team gets access without running the project locally. Lock down Studio with role-based access to define who can view, edit and execute. Bring in sales, support, compliance or QA to review traces, build datasets, run experiments and refine prompts, even if they don't have access to the source code.