This page describes what your AI agent can do with Chiplab.
You don’t need to call anything yourself; it’s reference material for the curious, not instructions to follow.
Chiplab’s exact tool names, parameters, and response shapes evolve as the platform grows.
Call ask with no arguments for the current, authoritative list — everything below is illustrative as of writing, not a fixed API contract.
Asking questions
Your agent can query Chiplab’s built-in knowledge base for platform, tool, and firmware guidance.
It’s purpose-built for Chiplab topics and won’t answer off-topic questions.
Every Chiplab session should start with a null/empty query, which returns a platform overview pulled straight from the knowledge base, listing the current, complete set of tools.
A natural-language query gets routed to an AI assistant scoped to Chiplab topics, grounded in the shared knowledge corpus.
{ "query": "How do I upload a firmware artifact?" }
{ "query": "What boards does Chiplab support?" }
Building, uploading, and running firmware
Your agent produces an ELF, uploads it, runs it on a chip-accurate virtual board, and reads back the result.
Chiplab runs your firmware on a virtual instance of the target board via a simulation platform (today that means Renode, with QEMU and more planned) and returns the console output, no physical hardware required.
Depending on what’s live when you connect, this might be a single call or a small sequence (e.g. issuing an upload slot, uploading the artifact, then triggering the run).
Your agent’s MCP client discovers the exact tool calls and parameters live; ask Chiplab if you want to see them yourself.
A typical run, illustrated as one step for simplicity:
{ "artifact_id": "a1b2c3d4-...", "board_id": "stm32f4_discovery" }
{ "run_id": "run-xyz", "stdout": "Hello from STM32F4!\n", "stderr": "" }
Not sure which board fits your firmware? Ask Chiplab about individual board specs; board recommendation tooling is on the way.