Routine care flows
Appointment confirmations, reminders, and PHI-free answers clear in well under a tenth of a second. Patients still get midnight answers.
For Healthcare & HIPAA
PHI blocked. Minimum necessary, enforced.
Your agents answer patients, handle records, and send reminders. GaaS checks every message before it leaves: PHI exposure is detected and blocked, minimum necessary is enforced, and every decision lands in an audit trail that's ready if anyone asks.
GaaS gives healthcare operators HIPAA-aligned control over AI agents: PHI is detected and blocked before a message leaves, minimum necessary is enforced on every disclosure, and a tamper-evident audit trail records every decision — HIPAA is one of 12 regulatory frameworks evaluated automatically.
org_demo · patient messages
Simulated
| Time | Agent | Action | Verdict | Latency |
|---|---|---|---|---|
| 09:41:02 | scheduler | send_sms → appointment_reminder | APPROVE | 36ms |
| 09:41:00 | patient-msgs | send_sms → lab_result_notice | ESCALATE | 51.8s |
| 09:40:56 | patient-msgs | send_sms → diagnosis_detail | BLOCK | 29ms |
| 09:40:54 | intake-agent | send_message → intake_form_link | APPROVE | 39ms |
| 09:40:50 | patient-msgs | send_message → care_instructions | MODIFY | 35ms |
| 09:40:48 | scheduler | send_sms → appt_time_location | APPROVE | 34ms |
| 09:40:44 | intake-agent | share_record → unverified_recipient | BLOCK | 35ms |
| 09:40:42 | scheduler | send_sms → reschedule_notice | APPROVE | 37ms |
GOVERNING
22,736 decisions
Simulated feed. Actual decisions include full audit records with hash-chain verification.
The Fear
One message with PHI in it is one too many.
The agent is genuinely useful: it answers patients at midnight, confirms appointments, follows up after visits. But every message it sends is a chance to say too much — a diagnosis in an SMS, a test result to the wrong thread, a detail beyond the minimum necessary.
A human assistant gets training, supervision, and consequences. Your agent got a paragraph in a prompt. And when someone asks "show me your safeguards," a prompt is not an answer. Hope is not a safeguard.
The Move
A safeguard that checks every message first.
GaaS sits between your agents and the patient. Before anything is sent, stored, or shared, it's checked:
Small fixes send corrected
A message that just needs a required disclaimer or safer wording is approved with that fix — corrected and sent, without waiting on a clinician.
Borderline messages wait
Anything touching results, diagnoses, or records is held for a clinician's yes before it goes anywhere.
PHI exposure never sends
Messages that would disclose PHI or exceed minimum necessary are stopped cold — before they leave, not discovered after.
Every decision lands in a tamper-evident audit trail — the answer to "show me your safeguards," already written.
Your Rules, In Your Words
Set policies the way you'd train a front desk.
No rule syntax, no config files. The conversational dashboard is powered by Claude — describe the boundary, and it holds:
"Never include a diagnosis or test result in an SMS."
"Hold any message that references lab results for clinician review."
"Appointment messages carry the date, time, and location — nothing else."
The Low-Friction Win
Wired in an afternoon. Zero risk to start.
Connect your agents
A developer wires the SDK in an afternoon — Python, TypeScript or Java, with plugins for LangChain, OpenAI Agents, CrewAI, Pydantic AI, Vercel AI, Microsoft Agent Framework, and MCP.
Watch in Shadow Mode
For the first week, GaaS just watches: it shows you exactly which messages it would have held or blocked, while changing nothing. Zero operational risk.
Turn it on
When the would-have-blocked list looks right, flip to live. From then on, the safeguard runs on every message, around the clock.
Pricing
Start free, stay cheap.
Start free in Shadow Mode, no credit card. There is a free tier, plans from $99 a month, and less than a cent per governed action at scale. Nonprofits, NGOs, and veteran-owned businesses govern free for life.
FAQ
Every objection, answered.
Does GaaS make my AI agent HIPAA compliant?
GaaS is the enforcement and evidence layer: it detects and blocks PHI exposure before a message leaves, enforces minimum-necessary disclosure on every action, evaluates decisions against HIPAA among its 12 built-in regulatory frameworks, and keeps a tamper-evident audit trail you can hand to an auditor. HIPAA compliance is broader than any one tool — but the two hardest parts, enforcement in real time and evidence on demand, are exactly what GaaS provides.
What is GaaS, in one sentence?
An external layer that checks what your AI agents are about to do and allows, fixes, holds, or blocks it against your rules, keeping a tamper-evident record of every decision.
Do I have to be a regulated business?
No. GaaS is for any operator running agents. Regulated teams get framework mappings; everyone else gets control over what their agents do.
Is it hard to set up?
No. Start in Shadow Mode with just an email; it runs the full pipeline on real actions without enforcing anything, so there is zero operational risk. A developer wires the SDK in an afternoon, and you author policies in plain language.
What does it cost?
Start free in Shadow Mode, no card. There is a free tier, then plans from $99 a month, and under a cent per governed action at scale. Nonprofits, NGOs, and veteran-owned businesses govern free for life. See pricing.
Will it slow my agents down?
Routine actions clear in well under a tenth of a second. Only high-stakes decisions take longer, and only because you asked them to.
Do I have to change my agent or my model?
No. GaaS sits outside the agent and needs no model changes and no cooperation from the agent to work.
Doesn't governance cost me tokens?
The opposite. A self-governing agent spends about 14,700 tokens of context on every routine governed action; with GaaS it spends 0 through a framework plugin (74 when an action is blocked), or about 2,470 through MCP. On a 200K-token model, self-governance fills 30% of the window after about 14 governed actions and 60% after about 28. See The Context Dividend.
I already have prompt guardrails. Why GaaS?
Prompt guardrails live inside the model, get re-read on every call, and can be argued away. GaaS is external and enforced; the agent cannot talk it out of a block.
Put the safeguard between your agents and your patients.
Start free in Shadow Mode — see exactly which of last week's messages would have been held or blocked, without changing a thing. No credit card, zero operational risk.
Not in healthcare? Find your door at GaaS for AI Operators.