The New Era of Private Data

DataHive AI released a new OpenClaw skill called Ride Receipts (DataHive Ride Insights). The skill extracts structured ride data from Gmail receipts for Uber, Bolt, Yandex, Lyft, Free Now, Curb, Via and other providers. All processing occurs locally on the user’s machine. It stores the data in SQLite and produces an anonymized CSV for optional upload to DataHive missions.

Technical Architecture and Data FlowThe skill uses this pipeline:

1. Email Ingestion The gog CLI fetches ride receipts from Gmail. Raw emails are saved todata/ride-insights/emails.json.
2. LLM Extraction The scriptextract_rides_gateway.py sends each email JSON (including HTML) to a local OpenClaw Gateway at /v1/responses. The LLM parses the receipt and outputs structured ride records to data/ride-insights/rides.json.
3. SQLite Storage The scriptinsert_rides_json_sqlite.py loads the extracted data into data/ride-insights/rides.sqlite. The database follows the schema in references/schema_rides.sql.
4. Anonymized Export The scriptexport_anonymized_rides_csv.py generates data/ride-insights/anonymized_rides.csv with only de-identified fields.

The skill requires a local OpenClaw Gateway (http://127.0.0.1:port or ::1) . It refuses any non-localhost URL.

Privacy Implementation

• Raw emails and full receipt JSON stay on the device.
• The anonymized CSV contains only: provider, email_month (YYYY-MM), start_time_15m, end_time_15m, currency, amount, distance_km, duration_min, pickup_city, pickup_country, dropoff_city, dropoff_country.
• No addresses, payment details, driver names, message IDs or raw email content are exported.
• Users review the report before any upload.
• The gateway enforces localhost-only connections.

OutputAfter installation the skill provides:

• Personal ride history in SQLite
• Total cost of rides in a single currency
• Summaries of spending, habits, repeated routes, anchor locations and time-of-day patterns
• Ready-to-upload anonymized CSV

Installation

Run the command: openclaw skills install ride-insights

Then start a new OpenClaw session and talk to the agent, or execute the bundled scripts directly:

• fetch_emails_json.py
• extract_rides_gateway.py
• insert_rides_json_sqlite.py
• export_anonymized_rides_csv.py

The skill is available on ClawHub:
Ride Receipts:https://clawhub.ai/datahiveai/datahive-ride-insights

A mission in the DataHive AI dashboard accepts the anonymized CSV for users who want to contribute to the shared data pool. The skill demonstrates local agent processing for personal data extraction and optional anonymized contribution to decentralized AI training data.