Python Library

DocuLayer ships as a first-class async Python library — use it inline in any application without running a separate MCP server.

Install

pip install doculayer

Quick Start

import asyncio
from doculayer import search, fetch

# Search live docs — returns verbatim sections with source attribution
results = asyncio.run(search("dependency injection", "fastapi"))
for r in results:
    print(r.score, r.section.title)
    print(r.section.source_url)     # "https://fastapi.tiangolo.com/tutorial/..."
    print(r.section.content)        # verbatim Markdown from the real docs
    print(r.section.cited_content)  # content with attribution header prepended

# Fetch a specific section
content = asyncio.run(fetch("httpx", section="AsyncClient"))
print(content)
# > **Source**: https://www.python-httpx.org/...
# > **Fetched**: 2s ago
#
# ## AsyncClient
# ...verbatim text...

API Reference

search(query, source, max_results=5)

async def search(
    query: str,
    source: str,
    max_results: int = 5,
) -> list[SearchResult]:

Searches live documentation for query within source. Returns a ranked list of SearchResult objects.

Parameters

Name	Type	Description
query	str	Keyword query — e.g. "field validators"
source	str	Package or URL identifier
max_results	int	Maximum sections to return

Returns — list[SearchResult]

Each SearchResult has:

result.score          # float — BM25 relevance score
result.section.title  # str   — section heading
result.section.source_url    # str — canonical URL
result.section.content       # str — verbatim Markdown text
result.section.cited_content # str — content with attribution prepended
result.section.trimmed(n)    # str — content trimmed to n words

---

fetch(identifier, section=None)

async def fetch(
    identifier: str,
    section: str | None = None,
) -> str:

Fetches documentation for identifier. Returns verbatim text with citation header.

Parameters

Name	Type	Description
identifier	str	Package or URL — same formats as search
section	str | None	Heading to extract; None returns the full page

Returns — str — verbatim Markdown with > **Source**: … header

---

Async Usage in Frameworks

FastAPI / asyncio native

from doculayer import search

@app.get("/docs-search")
async def docs_search(query: str, source: str):
    results = await search(query, source, max_results=3)
    return [
        {"title": r.section.title, "url": r.section.source_url, "score": r.score}
        for r in results
    ]

Jupyter / IPython

import asyncio
from doculayer import fetch

# Jupyter already runs an event loop — use await directly
content = await fetch("pydantic", section="Field")
print(content)

Synchronous wrapper

import asyncio
from doculayer import search

results = asyncio.run(search("streaming", "httpx"))

---

Core Types

DocSection

@dataclass
class DocSection:
    title: str          # heading text
    content: str        # verbatim Markdown body
    source_url: str     # canonical page URL
    depth: int          # heading level (1 = h1, 2 = h2, …)

    @property
    def cited_content(self) -> str: ...     # content with attribution header

    def trimmed(self, max_words: int) -> str: ...  # word-capped content

SearchResult

@dataclass
class SearchResult:
    section: DocSection
    score: float

FetchResult

@dataclass
class FetchResult:
    url: str
    sections: list[DocSection]
    fetched_at: float   # unix timestamp