Can we cap AI spend per class and per research group?
Yes. Budgets are set per team, which maps to a class, a lab, or a research group. Alerts fire before the limit so coordinators can adjust. Every call is logged.
Solutions · Universities
Did any student data leave, and who spent what?
Every class and research group gets its own budget, with alerts, and a shared workspace. Personal data is replaced with coded stand-ins such as PERSON_1 before the model sees the call, and every interaction produces a tamper-evident record your institution keeps.
Runs inside your own cloud account.
Last reviewed
Illustrative. No customer data.
The difference
Their way, and ours.
Their wayRaidu
One shared API key for everyone, no spend visibility.Budgets per class and per research group, with alerts before the limit.
Students paste personal details into consumer tools.Codes replace personal data before the model sees the call.
No record of what the model was asked or what it returned.Every call written once, on record, inside your own infrastructure.
Researchers pick from more than 60 AI model providers. Every call runs under your policies, with personal data replaced, and a tamper-evident record written to your own infrastructure.
Source: Tower AI Console, AI Firewall, and AI FinOps
What IT and faculty leadership ask
And what the system provides.
Can a class share notes, agents, and knowledge bases?
Yes. The AI Console lets a class or research group share a workspace: chat threads, agents built on prebuilt tools, and knowledge connectors. Every member’s calls run under the same policy set, with personal data replaced before the call reaches the model.
How do we keep student and research-subject data out of the model?
Raidu intercepts every call and replaces personal data with coded stand-ins before the model sees the prompt. The model receives the stand-ins, not the raw values. 79 personal-data and credential types are covered. Every replacement is logged.
Raidu capabilities
What each group gets.
AI Console
More than 60 AI model providers. Researchers pick the model. The institution sets the policy.
Read more ›AI FinOps
Budgets per class and research group, with alerts. Spend tracked to the team and the decision.
Read more ›AI Firewall
Personal data replaced with coded stand-ins before the model sees the call. 61 guardrail policies, 36 detection services.
Read more ›The product
See the screens.
Product windows from the films, with sample data. No customer data.
The record
The class call, on the record.
Every call a student or researcher makes produces a tamper-evident entry: team, model, what the model received (after stand-ins), and the policy verdict.
- Group
- Research group · NLP lab
- Model
- Budget remaining: $14.32
- Personal data
- Replaced with coded stand-ins
- Verdict
- Allowed and logged
- Proof
- Tamper-evident
sha256 16a99627ed0f2ae115e92cfd44a26575…
recomputed in your browser
published with the record16a99627ed0f2ae115e92cfd…
Hash matches
The call path
Where each call is checked.
- InterceptEvery console and API call from students and researchers
- DetectPersonal data found and replaced with coded stand-ins
- CheckBudget and policy for the class or group
- CallModel receives stand-ins only, within budget
- ProveWrite-once tamper-evident record, inside your infrastructure
Also covered
- Budget per class
- Budget per research group
- Coded stand-ins
- 60+ providers
- Write-once records
Common questions
Answered straight.
Can different classes have different model access?
Yes. Policy sets are configured per team. One class can be restricted to a specific model set; a research group can have access to the full list from more than 60 AI model providers.
What are coded stand-ins such as PERSON_1?
A label that stands in for a personal value, numbered so the context survives: a person becomes PERSON_1 and their record number MRN_1 everywhere in the conversation. The model can still follow who did what, but it does not receive the raw value.
Do students need separate accounts?
No. The AI Console is the institution’s deployment. Students and researchers access it under their existing credentials. The institution sets the policies; the call path handles the rest.
Can researchers use models from multiple providers?
Yes. The AI Console covers more than 60 AI model providers. Researchers pick the model their work needs. Every call goes through the same policy and personal-data checks.
Where are the records kept?
Inside your own cloud install. Records are write-once and kept in your infrastructure.