(Last Updated: March 10th, 2026)
Privacy matters here. The app is built around API access rather than consumer chat surfaces because that is generally the cleaner and more controllable route. Still, no one should pretend this turns AI into a vault. If something is extremely sensitive, the safest move is still not to paste it into any online AI system at all.
As of this policy date, the AI providers used by LM Council state that API data is not used to train their public models. You should still read their current policies for yourself, because provider policies can change.
We collect information in the following categories:
LM Council may process your instructions on a one-off or recurring basis when you create Tasks. That means the Service may re-run a saved workflow without you manually typing the same prompt again each time. Task runs create metadata such as schedule state, run state, credit estimates, logs, and artifacts.
Some features may allow you to connect external services. In the pre-Ltd rollout, GitHub App read-only beta is the main planned connected-source path. Later phases may include Gmail, Outlook, Slack, Notion, and Drive.
When you connect an external source, we may process connector metadata, approved scopes, selected resources, and source data needed to generate outputs. We do not treat all connected-source data the same way:
We use collected information to:
We share information in limited circumstances:
A current subprocessors list lives at /subprocessors.
This is one of the more important parts of the policy. For private connected-source data, the default design goal is: fetch what is needed, process it for the task run, persist the generated artifact and limited provenance metadata, and discard the raw private source content.
In plain English: we want the product value to come from the output, not from building a hidden mirror of your private repos, docs, or messages.
Tasks can multiply activity because they may run on a schedule, consume credits repeatedly, and create new artifacts from earlier source material. That is useful, but it also means automation can amplify mistakes faster than a one-off chat.
Retention defaults are currently intended as follows:
Some providers may apply their own retention policies to data processed through their APIs. ElevenLabs, for example, states that it may retain certain data by default to improve and secure its services.
User data is stored using Firebase and Google Cloud Platform with authentication controls, encrypted transport, and database or server-side access controls. We also use rate limits, logging, and access checks in API routes and worker functions.
No online system is perfectly secure. That is not a dramatic statement, just reality. We try to reduce risk, not pretend it can be made zero.
Where UK GDPR or similar laws apply, we generally process personal data on the basis of contractual necessity (to provide the Service), legitimate interests (to secure, improve, and operate the Service), consent (where you choose to connect third-party sources or enable optional processing), and legal obligation where required.
Your data, including connected-source content processed during task runs and chat sessions, may be transferred to and processed in countries outside the United Kingdom, including the United States. Our AI and infrastructure providers — such as OpenAI, Google, Anthropic, xAI, Perplexity, ElevenLabs, and GitHub — operate primarily in the US.
Where personal data is transferred outside the UK, we rely on the safeguards provided by our providers, which may include Standard Contractual Clauses (SCCs), the UK Extension to the EU-US Data Privacy Framework, or UK adequacy decisions, as applicable. You can request more information about the specific safeguards in place by contacting us.
You may have rights to access, correct, delete, restrict, or object to certain processing of your personal data.
If you revoke a connector or its scopes, tasks depending on that connector may pause, fail, or become blocked. Where possible, connector metadata and dependent resource references will be removed or marked inactive after revocation.
If and when LM Council enables Gmail or broader Google connector features, this policy will be updated with a dedicated Google API Data Use section covering exactly what is accessed, how it is used, what is stored, what is not stored, and how access is revoked.
We do not currently rely on the usual marketing-cookie circus. If that changes, the policy will change too. Until then, the no-cookie-banner aesthetic remains one of the few unambiguous wins of modern web life.
The Service is not intended for children and we do not knowingly collect personal information from children.
We may update this policy from time to time. Material changes will be announced through the Service or by other reasonable means.
Questions or concerns about privacy can be sent to support@lmcouncil.ai.