lineardb

What it does

LinearDB turns Linear into a local data model. It reads the API, validates the account and team boundary, writes a SQLite mirror, and lets agents query the result as data.

It is built for task analytics: team coverage, state mix, priority mix, assignee load, stale open work, labels, comments, attachments, history, and state spans.

Why it exists

Linear has tasks. Agents need data. The wrong browser profile or personal token can silently point at the wrong workspace, so LinearDB makes the credential profile and workspace check explicit before a sync runs.

Once the mirror exists, analytics are local. The agent does not need another API call to answer basic operational questions.

Install

pipx install git+https://github.com/eidos-agi/lineardb.git

Use

# Prove the credential sees the intended Linear team
lineardb --account greenmark auth-check --team-key GMW

# Mirror all visible teams into SQLite
lineardb --account greenmark sync --sqlite outputs/greenmark-linear.sqlite

# Analyze the local mirror without calling Linear
lineardb analytics --sqlite outputs/greenmark-linear.sqlite --team-key GMW

Data model

• Current state: teams, users, projects, labels, issue labels, issues.
• Time series: sync runs and issue snapshots.
• Related data: comments, attachments, issue history, and state spans.
• Identity: account profile, viewer, and Linear organization metadata.

Boundaries

• No Linear write mutations.
• No token persistence in repository files.
• Explicit account profiles fail closed. --account greenmark will not fall back to ambient credentials.
• SQLite dumps are local artifacts and should not be published unless the data owner approves that exact disclosure.

See also Foreman for delegated implementation work and Forges for the shipping standards behind this package.