Loading your workspace
Checking session state, onboarding progress, and live backend data.
Syncing calendars
Checking the latest availability and conflict data.
Preparing copilots
Loading the agent and chat context from the backend.
Verifying session
Confirming your access and onboarding status.
System-level documentation for the GraftAI orchestration track. Architecture, operational safety, and rollout conventions.
DOCS_LIVE_v3.0.82
// KERN_ARCH_v1
GraftAI runs as a high-performance monolith: FastAPI backend, Next.js frontend, async SQLAlchemy models, and worker-driven async jobs for non-blocking workflows.
FastAPI + SQLAlchemy async models
Public booking routes with signed action tokens
Worker cron for calendar and booking reminders
// ORCH_LOGIC_v3
Public booking supports availability calculation, conflict prevention, confirmations, reschedule, and cancellation with email lifecycle updates.
Timezone-aware slot generation
Tokenized attendee management links
Calendar event synchronization updates
// SEC_BASE_v2
Security headers, trusted host checks, auth token validation, and per-route rate limiting protect critical API workflows.
HSTS and strict frame policies
HMAC action token verification
Rate limits for booking endpoints
// EMBED_LOADER_v1
Embed booking directly into external websites through the GraftAI embed loader and dedicated iframe route.
Drop-in script: /graftai-embed.js
Autoload via data-graftai-embed
Direct username-based route support
For production readiness, ensure neural reminders remain enabled in workers, monitor synchronization delivery logs, and verify public action token flows after any kernel refactor.
// EMBED_PROMPT_v1.0
<div data-graftai-embed data-user="OPERATIVE_ID"></div> <script src="https://graftai.tech/relay.js" defer></script>