Which Google Docs does Lockstep read?

Only documents in folders you allowlist. It looks for requirement language in specs and one-pagers, the same extraction that runs on Notion, and every candidate needs a human confirmation before it binds.

Does Lockstep store the whole document?

No. It stores the extracted decision and an evidence quote anchored to its source, with secrets redacted at ingestion, not a copy of the doc.

Do we need a special template in Google Docs?

No. Lockstep reads the doc as written and looks for requirement language like 'must', 'never', and 'always'. Clear requirement sentences extract better, but there's nothing to adopt.

Lockstep for Google Docs

Lockstep reads allowlisted Google Docs so the decisions written into specs and one-pagers become an enforced record instead of dying in a doc no one re-reads.

What Lockstep captures from Google Docs

a constraint written into a spec doc
a decision captured in meeting notes
an edit that quietly changes the requirement

How the loop runs

Lockstep captures the decision from Google Docs, routes it to its owner for one-click confirmation that makes it binding, then briefs every engineer and every AI coding agent before they touch the code, and flags any pull request that drifts from it. See the full pipeline on how it works.

How to set up Lockstep for Google Docs

In the Lockstep dashboard, open Connections, pick Google Docs, and create the connection.
Allowlist the folders Lockstep should read, by folder ID or search term.
Share a spec or one-pager into an allowlisted folder. Lockstep scans it for requirement language.
Candidates land in the review queue with an evidence quote anchored to the doc. The owner confirms, edits, or rejects each one, and every capture, confirmation, edit, and rejection is logged with actor and timestamp.
Confirmed decisions brief every engineer and agent on the affected code, and the PR check flags drift.

Example: a decision captured from Google Docs

Illustrative example

In Google Docs

The "Billing retries v2" spec: "Failed card charges are retried at most three times over 7 days, then the account moves to past-due."

Drafted decision

Failed charges retry at most 3 times over 7 days, then the account is marked past-due. Evidence quote anchored to the spec.

Approved record

The spec's owner confirms it. An agent changing the retry job is briefed on the limit before it starts, and the evidence quote stays one click from the decision, so anyone can see where the rule came from.

Access & trust

Read-only, allowlisted documents. Contents redacted of secrets.

What Lockstep reads

Documents in the folders you allowlist

What Lockstep never does

Docs outside your allowlisted folders
Editing your documents
Making anything binding on its own: every candidate waits for a human to confirm it
Keeping a copy of the doc: Lockstep stores the extracted decision and an evidence quote, with secrets redacted at ingestion

Full details on the security page.

Frequently asked questions

Which Google Docs does Lockstep read?: Only documents in folders you allowlist. It looks for requirement language in specs and one-pagers, the same extraction that runs on Notion, and every candidate needs a human confirmation before it binds.
Does Lockstep store the whole document?: No. It stores the extracted decision and an evidence quote anchored to its source, with secrets redacted at ingestion, not a copy of the doc.
Do we need a special template in Google Docs?: No. Lockstep reads the doc as written and looks for requirement language like 'must', 'never', and 'always'. Clear requirement sentences extract better, but there's nothing to adopt.