How LitCo keeps your matter data confidential.
LitCo’s AI gateway requires zero data retention on every request it sends to an AI model provider. LitCo checks for that requirement on each request just before sending it, and a request without it does not go out.
LitCo does not use your firm’s content to train AI models.
Zero data retention on every gateway request
LitCo’s AI gateway requires zero data retention on every request it sends to an AI model provider. LitCo checks for that requirement on each request just before sending it, and a request without it does not go out.
Zero data retention means the provider does not keep the request or its response after answering, apart from a temporary cache, and does not use them to train models.
The service LitCo uses for first-pass document review also has zero data retention enabled for LitCo’s account.
Some work stays on LitCo’s servers.
LitCo builds the search index for your documents and transcribes dictation on servers LitCo runs. No outside AI model provider receives that work.
How LitCo stores your data
LitCo stores your firm’s matters, including Ana’s conversation history and working files, on servers LitCo runs with its cloud hosting provider and in backups of those servers. LitCo keeps this data until your firm asks LitCo to delete it. Closing an account ends sign-in and access; it does not by itself delete data. Some deletions in the app hide data from view rather than erase it right away, and backups keep copies for a limited period. Some records, such as audit and activity logs and support tickets, are kept longer for security, legal, and operational purposes.
Content your firm sends through Slack, email, or mail connect goes to those services under their terms.
Encrypted nightly database backups
LitCo backs up its database every night and encrypts each backup before it leaves the server.
An encryption key for each matter’s files.
LitCo encrypts each file stored for Matter A with key A.
LitCo ties each file to one matter. The key for one matter does not open files in another.
The key that wraps each matter’s file key stays inside a managed key service. LitCo’s servers send wrap and unwrap requests to that service but never hold the wrapping key.
Outside LitCo’s serversEnvelope encryption for each matter’s files
LitCo encrypts each matter’s stored files with AES-256-GCM. Each matter’s files have a separate key, and LitCo wraps that key with a key held in a managed key service. A production deployment refuses to use a key file on local disk. LitCo’s production database is stored on block storage that LitCo’s hosting provider encrypts at rest.
Row security in the database
Postgres row-level security is forced on the tables that hold matter data, and the application connects to the database as a role that is not exempt from row-level security.
Matter-based access control
LitCo limits what each person can reach by seat, role and matter membership, and limits what Ana can reach by matter. LitCo enforces ethical walls both when people read documents and when they share them.
A record of matter activity
LitCo keeps a step-by-step record of each Ana run. Database triggers make the matter activity log and import history append-only.
Share files with permissions set for each link.
A Files share link can require a password or a one-time code sent by email. You choose whether the recipient can view or download, and when the link expires. LitCo records each file a recipient opens, and confidential-view links watermark each copy.
Confidentiality at a glance.
The questions a security questionnaire asks, in the order they usually arrive.
| Model provider retention | LitCo’s AI gateway requires zero data retention on every request it sends to an AI model provider. LitCo checks for that requirement on each request just before sending it, and a request without it does not go out. The service LitCo uses for first-pass document review also has zero data retention enabled for LitCo’s account. |
| Model training | LitCo does not use your firm’s content to train AI models, and each gateway request tells the provider not to use it for training. |
| Work on LitCo’s servers | Building the search index and transcribing dictation, with no outside AI model provider. |
| What LitCo stores | Your firm’s matters, including Ana’s conversation history and working files, on servers LitCo runs with its cloud hosting provider and in backups of those servers. |
| Retention | LitCo keeps this data until your firm asks LitCo to delete it. Closing an account ends sign-in and access; it does not by itself delete data. Some deletions in the app hide data from view rather than erase it right away, and backups keep copies for a limited period. Some records, such as audit and activity logs and support tickets, are kept longer for security, legal, and operational purposes. |
| Integrations | Content your firm sends through Slack, email, or mail connect goes to those services under their terms. |
| Backups | LitCo backs up its database every night and encrypts each backup before it leaves the server. |
| Encryption at rest | AES-256-GCM on each matter’s stored files, with a separate key for each matter’s files. LitCo’s production database is stored on block storage that LitCo’s hosting provider encrypts at rest. |
| Key custody | The key for each matter’s files is wrapped by a key that stays inside a managed key service. A production deployment refuses to use a key file on local disk. |
| Access control | Seats, roles, and matter membership at the surface, with Postgres row-level security forced on the tables that hold matter data underneath. |
| Audit trail | Matter activity and import history are append-only, enforced by database triggers. |
For security questionnaires and diligence reviews, email hello@litco.ai. A list of LitCo’s service providers is available to customers on request.