Documentation
How to use Meshia, from launching your first GPU session to building autonomous research workflows.
New to Meshia?
Start with the step-by-step tutorial. It walks you from sign-in to a completed training run in about 8 minutes. Or press ⌘K anywhere in the app to find what you need.
Getting started
Tutorial: Your first experiment
Step-by-step walkthrough from sign-in to a completed training run. ~8 minutes.
Your first session
Sign in, pick a GPU, and chat with your agent in under 3 minutes.
The workspace
File explorer, terminal, GPU monitor, and experiments in one view.
SSH access
Connect from your own terminal with the SSH panel.
Core concepts
Agent roles
Hire a Researcher, Coder, Reviewer, Data, or Ops agent for your session.
Experiments
Track training runs with live metrics, compare results, reproduce configs.
Research plans
Let the agent build a hypothesis tree and execute experiments autonomously.
Canvas
A free-form surface where agents place charts, code, notes, and diagrams.
Workflows
Build multi-step DAGs of agent tasks and run them on a schedule.
Integrations
Platform
Billing and plans
Trial ($0, 1 concurrent session) and Pro ($500/mo, persistent workspaces and 3 concurrent sessions).
Agent memory
Your agent remembers context across sessions: preferences, findings, goals.
Multi-chat
Split your chat into multiple panes, each driving a different agent.
Keyboard shortcuts
Press ? anywhere to see every shortcut. Press ⌘K for the command palette.
Can't find what you need? Email us.