Welcome to
Silkroute Docs
Everything you need to build, deploy, and manage automation workflows that move your operations forward.
Overview
Silkroute automates back-office operations through AI-powered workflows. You describe what you want in plain English, connect your existing systems, and our agents handle the rest — from purchase order extraction to ERP sync, email routing, and exception handling.
This documentation covers everything from initial setup to advanced TRACE configuration. If you are new here, start with the Quickstart below.
These docs are for customers on an active Silkroute plan. If you do not yet have an account, reach out to sales@silkroutelabs.org.
Quickstart
You can have your first workflow running in under ten minutes. Here is the path:
Log into your workspace. Navigate to your Silkroute dashboard using the credentials from your onboarding email. Your workspace URL is app.silkroutelabs.org/<your-org>.
Connect an integration. Go to Settings → Integrations and connect at least one source (email inbox, ERP, or file storage). Silkroute supports 95+ integrations out of the box.
Create a workflow. Click "New Workflow" and describe what you want in the prompt box. For example: "When a purchase order arrives by email, extract the line items and push them to SAP."
Run in sandbox. Test your workflow against sample documents before activating it in production. Review the TRACE extraction output and adjust confidence thresholds if needed.
Activate. Flip the workflow to Live. Silkroute will begin processing in real time. Monitor run logs from the Activity tab.
Authentication
Silkroute uses API keys for programmatic access and OAuth 2.0 for integration connections. Your API keys are scoped to your organization and environment.
Go to Settings → API Keys → New Key. Assign it a name and select the permission scope. Keys are shown once — store them securely.
# Set your key as an environment variable export SILKROUTE_API_KEY="sk_live_xxxxxxxxxxxxxxxxxxxxxxxx" # Verify it works curl -H "Authorization: Bearer $SILKROUTE_API_KEY" \ https://api.silkroutelabs.org/v1/ping
| Scope | Access | Use case |
|---|---|---|
| workflows:read | Read | Fetch workflow definitions and run logs |
| workflows:write | Write | Create, update, activate, and delete workflows |
| extractions:run | Write | Submit documents to the TRACE engine |
| admin | Admin | Full access including key management and org settings |
Environments
Every Silkroute organization has two isolated environments: Sandbox and Production. Workflows, integrations, and API keys are scoped per environment.
Sandbox runs are free and use sample data. Switching to Production requires an active plan and connects to your live integrations.
# Sandbox base URL https://api.silkroutelabs.org/v1/sandbox/ # Production base URL https://api.silkroutelabs.org/v1/
Building Workflows
Workflows consist of a trigger, a sequence of actions, and optional condition branches. You can build them visually in the dashboard or define them as JSON via the API.
{ "name": "PO Intake", "trigger": { "type": "email_received", "filter": { "subject_contains": "Purchase Order" } }, "actions": [ { "type": "trace_extract", "schema": "purchase_order_v2" }, { "type": "erp_create_record", "integration": "sap_s4hana", "record_type": "purchase_order" } ] }
TRACE Engine
TRACE (Targeted Recognition and Classification Engine) is Silkroute's multi-model document extraction system. It reads PDFs, images, and email attachments and returns structured JSON fields with per-field confidence scores.
TRACE is the only extraction engine on the market that combines layout analysis, OCR, and semantic reasoning in a single pass — giving you higher accuracy on complex documents like multi-page POs, shipping manifests, and invoices.
curl -X POST https://api.silkroutelabs.org/v1/extract \ -H "Authorization: Bearer $SILKROUTE_API_KEY" \ -F "file=@purchase_order.pdf" \ -F "schema=purchase_order_v2"