docimprint 0.1.0

Python client and CrewAI integration for DocImprint verifiable document memory

docimprint (Python)

Verifiable document memory for AI agents — evidence bundles, Merkle proofs, and signatures.

Install

pip install docimprint           # REST client only
pip install "docimprint[crewai]"  # + CrewAI tools & toolkit (Python 3.10–3.13)

Set DOCIMPRINT_API_KEY or pass api_key= to the client. Default API base URL is https://api.docimprint.com.

CrewAI quickstart

import os
from crewai import Agent, Crew, Task
from docimprint.crewai import DocImprintToolkit

toolkit = DocImprintToolkit(
    api_key=os.environ["DOCIMPRINT_API_KEY"],
    collection_id="col_your_collection",  # optional; enables collection search/ask tools
)
researcher = Agent(role="Researcher", goal="Extract verifiable evidence", tools=toolkit.research_tools())
task = Task(description="Extract evidence from https://example.com/contract", agent=researcher)
Crew(agents=[researcher], tasks=[task]).kickoff()

DocImprintToolkit groups tools for research (ExtractEvidenceTool, SummarizeTool, QATool, CheckClaimsTool), legal (VerifyBundleTool, NotarizeTool), and collection search. Use toolkit.all_tools() for the full set.

REST client

from docimprint import DocImprintClient

with DocImprintClient(api_key="...") as client:
    bundle = client.extract_evidence(url="https://example.com")
    print(bundle["bundle_id"])

Docs

• Full CrewAI spec: docs/crewai-integration.md in the DocImprint SDK repo
• API reference: https://docimprint.com/docs

License

MIT — see LICENSE.