Meshia Docs

Documentation

Practical guides for launching compute, using the workspace, tracking experiments, and understanding the current platform surface.

Tutorial: Your first experiment

A guided 8-step walkthrough from sign-in to a completed training run. Takes about 8 minutes.

Getting started

Your first session

The workspace

Understand the current session workspace: files, research, compute, terminal, SSH, and activity.

SSH access

Generate a workspace key and connect to live compute from your own terminal.

Core concepts

Agent roles

Coming soon

Specialized agent roles are being wired into the workspace UI.

Experiments

Track training runs, metrics, artifacts, and comparisons from the research workspace.

Research plans

Use the research flow to turn a goal into structured hypotheses and runs.

Canvas

Coming soon

Spatial research surfaces are being prepared for the main workspace.

Workflows

Coming soon

Repeatable multi-step agent pipelines are under active buildout.

Integrations

GitHub

Connect a repository so the agent can work in a session-scoped checkout.

Webhooks

Coming soon

Webhook delivery foundations are in place; the public settings UI is coming soon.

Notifications

Configure the product notifications that are live in Settings.

Platform

Billing and plans

Plans, usage billing, and the billing settings flow.

Agent memory

How persistent agent pointers work across sessions.

Multi-chat

Split active work across chat panes without losing session context.

Keyboard shortcuts

Current keyboard shortcuts and command palette behavior.

Reference

GPU catalog

Available GPU catalog, fallback rates, capacity notes, and picking guidance.

Agent tools

A practical guide to the tool families surfaced in agent logs.

Security

Data isolation, authentication, secrets, Git safety, and retention.

Limits and quotas

Current quota, runtime, API, pod, and storage limits.