Knowledge Bases - ChatATP Studio Documentation

Use client.knowledge_bases to create, populate, search, and attach knowledge bases to Agents.

Create a knowledge base

from chatatp_studio import ChatATPClient

async with ChatATPClient(api_key="chatatp_sk_...") as client:
    knowledge_base = await client.knowledge_bases.create(
        "Product docs",
        description="Support documentation",
        agent_id=7,  # optional: attach while creating
    )

Upload and monitor documents

await client.knowledge_bases.upload(
    knowledge_base["id"],
    "./manual.pdf",
)

documents = await client.knowledge_bases.documents(knowledge_base["id"])
stats = await client.knowledge_bases.stats(knowledge_base["id"])

for document in documents["data"]:
    print(document["name"], document["status"], document["chunk_count"])

Indexing is asynchronous. Poll documents until their status is indexed or failed. Document responses include chunk_count, char_count, and error_message.

Index URLs and search

await client.knowledge_bases.add_url(
    knowledge_base["id"],
    "https://docs.example.com",
)

result = await client.knowledge_bases.search(
    knowledge_base["id"],
    "How do I reset my password?",
    top_k=5,
)
print(result["context"], result["has_context"])

Attach to an Agent

await client.knowledge_bases.attach(
    7,
    knowledge_base["id"],
    enabled=True,
    auto_context=True,
    max_context_chunks=5,
    context_threshold=0.3,
)

For the equivalent Studio CLI workflow, see the Knowledge Base section in the CLI overview.