Legal
AI Usage Policy
Updated 27 October 2026. Applies to TRYSA AI, the Credit Centre and any AI-powered feature.
How TRYSA AI works
TRYSA AI runs on open-source language models operated on TRYSA's own Cloudflare account. We do not send your prompts, conversations or workspace records to any third-party AI provider. When you ask a question, TRYSA looks up the records in your workspace that are most relevant, adds live figures for things like client balances, unpaid invoices and upcoming bookings, and gives them to the model so its answer reflects your own data.
What it can see
- Only your own workspace, and only the areas your role is allowed to view. If you cannot open invoices, TRYSA AI cannot read them for you.
- Clients and leads, bookings, invoices, contracts, projects, tasks, services, campaigns, gallery content and Academy lessons.
- It does not read passwords, API keys, private direct messages or payment proofs.
- To find relevant records, TRYSA stores searchable copies of their text in a private index. Each copy is tagged to its workspace and is never used for another workspace.
Your data and training
We do not use your workspace content to train or fine-tune any model, ours or anyone else's.
Credits and metering
Every AI request is server-metered against your Credit Centre balance. Credits are based on the length of the question and the answer and are recorded in an append-only credit ledger. Workspace admins can suspend AI access.
Accuracy and your responsibility
TRYSA AI can be wrong. It can misread a record, miss one, or state something confidently that is not true. Check anything that matters — especially amounts, dates, invoices, contracts and client commitments — before you act on it or send it to a client. You are responsible for what you publish or send.
Content ownership
You own the prompts you write and the outputs generated on your behalf. Outputs may resemble other outputs because they come from the same models, so we do not promise uniqueness.
Acceptable use
Do not use TRYSA AI to generate illegal content, disinformation, targeted harassment, malware or deceptive content, to impersonate others, or to attempt to extract other workspaces' data or bypass limits. We may restrict access for misuse. See our Acceptable Use Policy.
Reliability
TRYSA AI depends on our infrastructure and can be slow or unavailable at times. We do not guarantee uninterrupted service.
