Consuming artifacts

Reference for using Podium artifacts. Each page below covers one topic.

Page What it covers
Configure your harness Per-harness setup for podium sync (filesystem materialization) and the MCP server (runtime discovery). Covers Claude Code, Claude Desktop, Cursor, OpenCode, Codex, Gemini, Pi, Hermes, and the generic/none adapter.
Browsing the catalog How an agent navigates the catalog at runtime: load_domain, search_domains, search_artifacts, load_artifact. The discovery flow and what each call costs.
Custom consumers via the SDK Building programmatic consumers (LangChain, Bedrock, OpenAI Assistants, custom orchestrators, eval harnesses) with podium-py or podium-ts.
Handling artifact responses What a consumer does with the manifest and materialized files returned by load_artifact: which bytes arrive in the response and which are written to disk, routing by hints, safety and trust constraints, capability checks, dependency walks, and external resources.

Pick the page that matches the consumer under configuration.


Table of contents