Terms and Conditions — BrainOutput trial workspace
Read me first. A trial workspace is a computer that runs software and AI
models for you, on infrastructure we operate in the European Union. These terms tell you what it is,
what you may do with it, what we do with your data, and where the limits are. They are written to be
readable; where something is ambiguous, the stricter reading for us applies.
1. What the service is
BrainOutput provides a hosted, isolated workspace running the BrainOutput Community Edition:
a chat-based "AI company" that can read sources you connect, draft and produce work, and execute
tasks on AI models you select (free third-party models, your own provider keys, or local models).
The trial lasts seven (7) days from creation, requires no payment card, and is intended for
evaluation only.
2. Your account and acceptance
- One workspace per person, subject to anti-abuse limits (currently three workspaces per day per address or network).
- You must provide a real email address you control and explicitly accept these Terms at creation time. Your acceptance is recorded with its version and date.
- Your workspace link contains an access token. It is shown to you once. Anyone with the link can use the workspace — keep it private. We cannot email it to you again.
3. Acceptable use — the hard rules
You may not use the workspace, directly or through an AI agent, to:
- Break the law, or help anyone else break it — in any jurisdiction that applies to you or to us (including French and EU law);
- Attack, scan, probe, or disrupt any network, system, or account (ours or anyone's);
- Send unsolicited bulk messages (spam), or produce deceptive content at scale;
- Mine cryptocurrency or otherwise consume resources for purposes unrelated to evaluating the product;
- Process personal data of third parties without a lawful basis, or upload content you have no right to use;
- Attempt to escape the workspace's isolation, access other tenants' data, or interfere with other users;
- Misrepresent AI-generated output as human-written where that would be deceptive, or use it to impersonate a real person without consent.
We may suspend or delete a workspace immediately for a breach of this section, without notice
if the situation requires it (e.g. active abuse), and report to authorities where the law obliges us to.
4. AI-generated output — what you must understand
- Output is produced by third-party or local AI models. It can be wrong, incomplete, fabricated, or infringing. You are responsible for reviewing anything before you use, send, or publish it.
- Free third-party models (e.g. currently-free providers) may, under their own terms, use prompts to improve their models during the free period. Do not send confidential or personal data through free models. For sensitive work, use your own provider key or a local model.
- The system is designed to ask for human approval before consequential actions, but it is not a guarantee: configuration errors or model behavior can produce unintended output. Approval gates reduce risk; they do not eliminate it.
5. Your data and our data practices (GDPR)
- What we store: your email address (as account identifier and for abuse prevention), your terms acceptance record, and the workspace's content (which you control). Technical logs (IP, requests) are kept for security and abuse investigation.
- Legal basis: contract performance (providing the workspace), legitimate interest (security, abuse prevention), and compliance with legal obligations.
- Processors/sub-processors: hosting is provided by EU-located providers; AI inference goes only to the model providers you select (free providers, your own keys, or local models). Your workspace content is never sold and never used to train models by us.
- Retention: workspace content is deleted after expiry, except where you ask us to migrate it (upgrade path) or the law requires otherwise. The email + acceptance record is kept while needed for abuse prevention.
- Your rights (GDPR): access, rectification, erasure, restriction, portability, and objection — write to contact@brainoutput.com. You also have the right to lodge a complaint with the CNIL (France) or your local authority.
6. Availability and the trial's limits
- The trial is provided "as is", without any service-level commitment. It may be paused, rate-limited, or stopped for maintenance, capacity, or abuse reasons.
- Workspaces are isolated (rootless containers, restricted networking), but no isolation is absolute: you accept the residual risk inherent to shared-kernel infrastructure for a free evaluation tier.
7. Liability
To the maximum extent permitted by applicable law, BrainOutput is not liable for indirect or
consequential damages, lost profits, lost data, or business interruption arising from the trial.
Nothing in these terms excludes liability that cannot be excluded by law (including liability for
fraud, gross negligence, or death/personal injury caused by negligence).
8. Changes and law
- We may update these terms; the version you accepted is recorded, and material changes require a new acceptance for continued use.
- These terms are governed by French law. Courts of France have jurisdiction, without prejudice to mandatory consumer-protection rules in your country of residence.