# Frequently Asked Questions

## What is DocuSOR?

DocuSOR is shared, verified context and memory for coding agents. It writes a baseline of facts
about each repository, proves them against your running app in an isolated sandbox, and serves
them through the DocuSOR MCP server to every agent and every engineer on your team, whichever tool
or model they use. Agents start each task already knowing how your app works instead of
re-exploring the repo. See the [documentation](https://docusor.ai/docs).

## Is it tied to one coding agent or model?

No. Context and memory live in DocuSOR, not inside any one tool or model. Claude Code, Cursor,
GitHub Copilot, Codex, Windsurf, Cline, and any agent that can connect to a remote MCP server over
HTTP read the same verified facts. Switch or mix agents and nothing is lost.

## How is this different from CLAUDE.md, Copilot memory, or Cursor rules?

Those files live inside one tool and often one person's setup, and nothing checks that they are
still true. DocuSOR is one memory shared by every agent and engineer, and every fact has a check
that runs against your app. Context whose checks fail or drift is marked stale. In a Claude Opus
5.5 benchmark against an agent using its own memory, DocuSOR matched quality while finishing 77%
faster at 64% lower model cost. See the [case study](https://docusor.ai/case-studies).

## How do I get started?

Sign in, then connect your repositories: install the DocuSOR GitHub App on the repositories you
want covered, or, on Enterprise, connect GitLab (cloud or self-managed). DocuSOR onboards each
one automatically, for free. Then connect your coding agent to the DocuSOR
MCP server. The [Getting started guide](https://docusor.ai/docs#getting-started) has the steps, and
[MCP setup](https://docusor.ai/docs#connect-your-coding-agent-mcp-setup) has configs for Claude Code, Cursor,
VS Code, and Windsurf.

## What does DocuSOR validate?

The context it keeps for your repository: the baseline written during onboarding plus anything your
agents add. Each fact has a check, and DocuSOR runs those checks against your app in a sandbox.
The context lives in DocuSOR, not in a file in your repository.

## What does my repository need?

A Compose file lets DocuSOR start your app automatically. Without one (including Dockerfile-only
repositories), DocuSOR doesn't start the app, so endpoint checks need a start step that your
agent can add. See [What your repository needs](https://docusor.ai/docs#what-your-repository-needs).

## Can DocuSOR test my README and runbooks?

Yes. Write checkable facts as short claims, such as "`GET /health` returns `200`", then click
**Test a doc** on the repository page. DocuSOR reports each claim as passed or failed. See
[Writing testable docs](https://docusor.ai/guides/testable-docs).

## How does billing work?

Standard uses prepaid credits at **$0.10 each** (as low as $0.08 with larger packs), and new
organizations get 25 free credits. One credit covers one context read by an agent (such as
`get_repo_context` or `query_context`) or one run an agent starts. Onboarding, scheduled
re-validation, and anything you start from the dashboard are free. Enterprise adds an annual
platform fee with prepaid credits. See [Credits and billing](https://docusor.ai/docs#credits-and-billing).

## What does Enterprise include?

Unlimited repositories and seats, GitHub and GitLab (GitLab.com or self-managed), SSO over SAML
or OIDC, teams with per-repository access, self-hosting in your own infrastructure (including
air-gapped), private runners as an add-on, audit logs, and advanced support. Pricing is an annual
platform fee plus prepaid credits. See [Enterprise](https://docusor.ai/docs#enterprise).

## What does it cost to test a doc?

Testing a doc from the dashboard is free. An agent testing a doc with `check_doc` uses one
credit, and only if the doc contains at least one claim that runs. Otherwise it is refunded.

## Does testing a doc change my repository's status?

No. A doc test only reports on that document. It never changes whether your repository shows as
"Validated."

## Does it work with GitHub and GitLab?

Yes. Standard plans connect GitHub, and Enterprise plans connect both GitHub and GitLab. On
GitHub, the DocuSOR App has read-only access to Contents and Metadata and no write permissions.
On GitLab.com or self-managed GitLab,
you connect a token with `read_api` and `read_repository`, or `api` if you want
results posted as commit statuses. Uninstall the App or revoke the token at any time. See
[Repository access](https://docusor.ai/docs#repository-access).

## Does DocuSOR store my source code?

No. Code is fetched from GitHub or GitLab for each run and the sandbox is destroyed when the run
ends. DocuSOR keeps the verified context for your repository and your run history: status,
timing, and a report of each check's result, including the output the check produced.

## Where do my runs execute?

On DocuSOR's managed infrastructure, in isolated containers that are destroyed when each run
ends, whether it passes or fails. A run executes your repository at a specific commit plus the
checks stored in DocuSOR. Only people and agents with your organization's keys can add checks.
`validate_changes` also includes the uncommitted files an agent submits.

## How is my data protected?

Data is encrypted in transit with TLS and at rest. Production infrastructure sits behind a
default-deny firewall, and the database runs on private networking with no public exposure.
GitLab tokens are encrypted at rest. See [Security](https://docusor.ai/security).

## How do I report a security issue?

Email security@docusor.ai. We'll acknowledge your report and work with you on a fix. Please give
us a reasonable chance to address it before public disclosure.

## Where can I see product updates or system status?

Updates are on the [Changelog](https://docusor.ai/changelog). Live status is on the [Status page](https://docusor.ai/status).
