Commit Graph

2 Commits

Author SHA1 Message Date
Khalid Abdi b29fdff1cb feat: ambient AI visit scribe (record/paste → reviewed SOAP note)
Record a clinician↔patient visit (or paste a transcript) on the patient
sheet; the backend transcribes it (OpenAI Whisper / Gemini), de-identifies
the transcript + context through Veil, and drafts a structured SOAP note
the clinician reviews and edits before saving — the same write-approval
gate as the chat agent.

Backend: POST /api/scribe/{transcribe,draft,save} (routes/scribe.ts,
services/ai/transcribe.ts), veil.redactText() free-text redactor,
appendEncounter service, audio MIME types on attachments. Gated by
patient:write + the clinic AI policy (reception/disabled-AI excluded).
Frontend: ScribeDialog + lib/scribe.ts, "Record visit" on the patient
detail, gated by clinical access + AI availability. New `scribe` locale
namespace across all five languages. Bumps to v0.4.0.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-03 18:33:30 +03:00
Khalid Abdi 3a7378e00d backend: AI chat agent with Veil PHI safeguard (providers, /chat, import)
Real LLM chat replacing the mock, backend-centric per the plan:

- Multi-provider API-key mode (OpenAI / Anthropic / Gemini via the AI SDK) plus
  local Ollama (OpenAI-compatible endpoint). Provider is derived from the
  picked model id; the matching stored key is used. New user_ai_settings table
  holds per-user config with provider API keys encrypted at rest (AES-256-GCM,
  src/lib/crypto.ts, keyed by AI_CREDENTIALS_KEY).
- POST /api/ai/config (get/put, secrets never returned), POST /api/ai/test
  (Ollama ping / key presence), POST /api/ai/import (approved migration commit,
  re-validated server-side, reuses the audited patient service).
- POST /api/chat: streamText agent with tools (getPatient, getPatientLabs,
  searchPatients, previewImport). Real record data streams to the clinician as
  custom data parts (cards) while the model sees only Veil-redacted results.
- Veil (src/services/ai/veil.ts): de-identifies patient identifiers to tokens
  before external calls, resolves tokens on tool args, and rehydrates the final
  answer. Bypassed for local Ollama. External mode runs non-streamed so the
  rehydrated text is correct. Every call is audited (provider + Veil level).
- Shared role-scoping helpers extracted to src/lib/role-scope.ts (reused by the
  patient routes and chat tools so visibility rules match).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-13 18:35:16 +03:00