> ## Documentation Index > Fetch the complete documentation index at: https://docs.veecle.ai/llms.txt > Use this file to discover all available pages before exploring further. # Core concepts > The few ideas behind every Chiplab session: projects, boards, artifacts, runs, and the knowledge corpus. ## Project A project is where your work lives. Uploads, builds, and runs stay in the project, so your agent can pick up where it left off next session. ## Board A board is the virtual board to run on: a board model plus the UARTs to capture. Only the UARTs you list are captured. Models and UART names are on [Hardware](/hardware/boards). ## Artifact An artifact is the compiled firmware (ELF) your agent uploads. Chiplab runs it as-is. ## Run A run is one execution of an artifact on a board. It lasts up to 5 seconds of virtual time by default and ends early if the firmware finishes or panics. The result is a status and the captured output. ## Knowledge corpus The knowledge corpus is what Chiplab learns about how each chip behaves. It holds chip-level behavior only, not your source code or binaries. Your agent draws on it through [ask](/tools/ask). The longer description is on the [introduction](/introduction#knowledge-corpus). ## Your agent and ask You describe the work. Your agent calls the tools. [ask](/tools/ask) is how it learns how Chiplab works. Run your first simulation in under five minutes. What ask, build, run, test, and CI/CD each do. This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.