llm-api-scope 0.8.0

read and cache structured documents from remote for LLM agents

LLM API Scope (apiscope)

a tool for LLM agents to read and cache structured documents from remote.

install

use pipx for isolated installation:

pipx install llm-api-scope

how it works

apiscope fetches specifications from local files or remote URLs, caches them, and outputs structured JSON that agents can consume directly. no html parsing, no keyword ranking — just faithful extraction from the source document.

commands

run apiscope --help to see all available commands.

openapi

browse OpenAPI specifications with subcommands for discovering, listing, and describing operations.

aliases let you register frequently used specs once and reference them by short name. fetching is transparent — local copies are cached for fast repeat access, and a proxy can be configured for restricted networks.

rfc

read, search, and navigate RFC documents from the IETF.

the metadata index is mirrored once via rsync. individual text files are fetched on demand and cached locally. you can browse the table of contents, jump to a section (XML) or page (TXT), filter the index by status or source, and run keyword searches against fulltext content.

skill

output a combined command reference and strategy guide for AI agents.

run apiscope skill to get a single document that covers every subcommand and how they fit together. the strategy guide classifies commands into management and usage groups, lists prerequisites in ABNF (RFC 5234), and provides workflow patterns with requirement levels (RFC 2119).

agents read this output once at onboarding instead of running --help repeatedly.

repo

sync documentation directories from any git repository to a local cache.

register repositories by URL with a target subdirectory and a ref — a branch, tag, or commit. the sync command clones with shallow depth, blobless filter, and sparse checkout so only the needed tree and files come over the wire. extracted docs land in the cache under a stable hash path, and a configurable TTL avoids redundant re-fetches. provider-agnostic, zero auth.

future

• read academic papers from arxiv
• more formal document formats as the need arises

license

MIT