Load CVAT data to DuckDB

CVAT is a web-based tool for annotating images and videos that provides a REST API for programmatic interaction and management of annotation projects and tasks. Everything needed to build a working CVAT → DuckDB pipeline is on this page: the API's base URL, authentication, endpoints, pagination and incremental field — plus a prompt that hands the whole job to your coding agent.

---

Build your CVAT to DuckDB pipeline

Paste this prompt into Claude, Codex, or Cursor. The agent does the rest.

PromptCopy prompt
Run uvx dlthub-init@latest to build a pipeline from CVAT to DuckDB and run it on dltHub

That scaffolds a dltHub workspace and installs the dltHub AI harness — the project rules, the secrets-management skill, and the dlt MCP server your agent needs to work safely. From there it reads the CVAT API, proposes the endpoints to load, then writes, runs and validates the pipeline while you review rather than type. Credentials are inspected through MCP tools, so your agent never reads secrets.toml itself. How the LLM-native workflow works →

Prefer to write it yourself? Every fact the agent uses is below.

---

CVAT API at a glance

These values come from the CVAT API documentation. Check them against the vendor's current reference before relying on them in production.

---

How do I authenticate with the CVAT API?

CVAT API requests require an Authorization header. For Personal Access Tokens (PAT), use the format 'Authorization: Bearer <token_value>', while for legacy Token authentication, use 'Authorization: Token <token_value>'.

1. Get your credentials

To obtain a Personal Access Token (PAT) for the CVAT REST API, follow these steps in the web interface: 1. Log in to your CVAT instance. 2. Navigate to your Profile page. 3. Go to the "Security" section. 4. Click the "+" button to create a new token. 5. Configure the token name, expiration date, and desired permissions (read-only or read/write). 6. Save the token and store it securely, as it is only displayed once. PATs are the recommended, secure authentication method for programmatic access.

2. Add them to .dlt/secrets.toml

Copy
[sources.cvat_source]
cvat_api_token = "your_personal_access_token_here"

dlt reads this file automatically at runtime. With the harness, the setup-secrets skill prompts you for the values and never handles the raw credential in chat. For production, see setting up credentials with dlt.

---

What CVAT data can I load into DuckDB?

These are the CVAT endpoints dlt can load into DuckDB:

---

How do I load only new CVAT records?

CVAT exposes page on api/tasks, so dlt can request only the records that changed since the last run. Set it as the cursor_path and dlt tracks the high-water mark for you between runs.

Copy
{"name": "tasks", "endpoint": {
    "path": "api/tasks",
    "data_selector": "results",
    "incremental": {"cursor_path": "page", "initial_value": "2024-01-01T00:00:00Z"},
}}

On the first run dlt loads everything from initial_value; on every run after that it requests only what changed and appends with write_disposition="merge" if you set a primary key. See incremental loading.

---

What does the generated CVAT pipeline look like?

A standard dlt REST API pipeline — the same code you would write by hand, loading tasks and projects from the CVAT API into DuckDB:

Copy
import dlt
from dlt.sources.rest_api import RESTAPIConfig, rest_api_resources

@dlt.source
def cvat_source(api_key=dlt.secrets.value):
    config: RESTAPIConfig = {
        "client": {
            "base_url": "https://app.cvat.ai/api",
            "auth": {"type": "bearer", "token": api_key},
        },
        "resources": [
            {"name": "tasks", "endpoint": {"path": "api/tasks", "data_selector": "results"}},
            {"name": "projects", "endpoint": {"path": "api/projects", "data_selector": "results"}}
        ],
    }
    yield from rest_api_resources(config)


def load_cvat_to_duckdb() -> None:
    pipeline = dlt.pipeline(
        pipeline_name="cvat_pipeline",
        destination="duckdb",
        dataset_name="cvat_data",
    )
    load_info = pipeline.run(cvat_source())
    print(load_info)


if __name__ == "__main__":
    load_cvat_to_duckdb()

Run it with python cvat_pipeline.py. The agent iterates on this until it loads cleanly — you review and approve, rather than write it from scratch.

---

How do I query CVAT data in DuckDB?

dlt creates one table per resource. Query the loaded data with Python or SQL — or ask your agent to, through the MCP server's execute_sql_query tool.

Python (pandas DataFrame):

Copy
import dlt

data = dlt.pipeline("cvat_pipeline").dataset()
df = data.tasks.df()
print(df.head())

SQL:

Copy
SELECT * FROM cvat_data.tasks LIMIT 10;

See querying your data with dataset and exploring it in marimo notebooks.

---

How do I deploy the CVAT to DuckDB pipeline in production?

The pipeline runs locally, which is ideal for prototyping and one-off analysis. When you need it on a schedule, monitored on every load, and shared with your team, deploy the same dlt code on the dltHub platform — no infrastructure to maintain. The prompt above already ends with "run it on dltHub", so your agent can take it there directly.

• Deploy & schedule — run the pipeline as a managed job with automatic retries.
• Monitor — observable job queues, alerting, and load metrics for every run.
• Transform — promote raw CVAT loads into governed, documented models.
• Visualize & share — explore data in notebooks and publish live dashboards instead of static screenshots.

Book a demo →

---

What other destinations can I load CVAT data to?

dlt loads into any of these — only the destination argument changes:

Set dlt.pipeline(destination="snowflake") and add credentials in .dlt/secrets.toml. On the dltHub platform the same pipeline runs against a managed Iceberg lakehouse. See the full destinations list.

---

Next steps

• Ready for production? Book a demo of the dltHub platform →
• dlt LLM-native workflow docs
• REST API source reference
• Join the dlt community on Slack