Quickstart¶
Hydron is a hardware-native AI coding assistant for embedded, robotics, and UAV engineers. It runs as a VS Code extension and as a terminal CLI, both sharing the same context engine, agent modes, and hardware integrations.
This guide takes you from a signed-in client to your first hardware-grounded answer. Your administrator provides the extension package or CLI installer.
Before you begin¶
Hydron grounds its answers in your specific hardware and software constraints, so it needs real context:
- An active development codebase.
- A configured project. See Project Dashboard.
Step 1: Authenticate¶
Open the Hydron chat pane and click Login with Hydron. The sign-in page opens in your browser and shows Authentication Complete when done.
Step 2: Select your project¶
Your project carries your platform documentation, schematics, and indexed codebase. Every answer is scoped to it.
Click the project dropdown at the bottom of the chat pane and select your project.
Press Ctrl+Shift+P to open the command palette and select your project.
If no project exists yet, set one up first. See Project Dashboard.
Step 3: Index and ask¶
Before Hydron can write or explain firmware, it needs to see your local files.
Hydron scans your workspace, maps your directory structure, and creates a hydron.md file at your project root. That file holds the project's baseline rules, and Hydron reads it on every session from then on.
Edit hydron.md to record conventions the agent should always follow, and commit it so your team shares the same baseline.
Then ask:
That is your first hardware-aware answer.
Tip
To ask about a single file without indexing the whole project, type @ followed by the filename, for example @src/main.c, to pull it directly into the chat.
Step 4: Ground the answer in your hardware¶
Generic answers come from training data. Grounded answers come from your attached datasources.
Query the DRV8316 datasheet in the project and check the register write
sequence in @src/drivers/drv8316.c against it.
Use trigger words such as query, refer to, or look up so Hydron routes to your datasources rather than answering from training data. See Best practices.
Next¶
- Project Dashboard to give Hydron your platform docs, schematics, and codebase.
- Agent modes to pick the right mode for a task.
- Best practices for the patterns that make answers accurate.