zizkadb-sdk

Python SDK for ZizkaDB — the operational database for AI agents.

Setup (managed cloud)

1. Sign up at db.zizka.ai
2. Dashboard → Create agent (e.g. my-bot) — you get an API key for that agent
3. Use the same agent name in every db.log() call
4. Set ZIZKADB_API_KEY (or pass the key to the constructor)

Important: The agent in db.log(agent="...", ...) must match the agent you created in the dashboard. A mismatch returns 403 AgentScopeError.

Install

pip install zizkadb-sdk

Note: There is an unrelated package called agentdb on PyPI. Install zizkadb-sdk. Import: from zizkadb import ZizkaDB.

Scaffold a project

pip install zizkadb-sdk
zizkadb init my-agent --template basic
cd my-agent && cp .env.example .env   # paste your key into ZIZKADB_API_KEY
pip install -r requirements.txt
python agent.py

Templates: basic, openai, langchain, crewai, mcp-cursor.

Quickstart

from zizkadb import ZizkaDB

# Key from Dashboard → Settings → API keys → create "my-bot" → copy key
db = ZizkaDB("zizkadb_live_xxxx")

async with db:
    result = await db.log(
        agent="my-bot",  # must match dashboard agent name
        event="tool_call",
        data={"tool": "search", "query": "billing"},
    )
    chain = await db.why(result.event_id)
    chain.print()

Multi-agent apps (one key, many agent names)

If your app logs to different agent ids per user (e.g. conv-alice, conv-bob), create a tenant-wide key in Settings → Tenant-wide API key. Per-agent keys only work for that one agent name.

Errors

Environment variables

Self-host dashboard

1. bash scripts/setup-local.sh (API + dashboard on port 3001)
2. Open http://localhost:3001/login → Open my dashboard →

Links

• Docs
• API explorer
• GitHub