Documentation Portal

Welcome to the Deep Retrieval Infrastructure (DRI) developer portal. This is a holding page outlining the core system architecture and query standards.

System Architecture

DRI sits beneath the application layer as a federated retrieval gateway. Instead of crawlers indexing the general web or agents writing custom scraping scripts, DRI translates query payloads into targeted, parallel requests routed to official database registries.

Quickstart

Query the federated index using a standard JSON request POSTed to our endpoint. Ensure you authenticate using your developer API key:

curl -X POST https://api.getdeep.tech/v1/query \
  -H "Authorization: Bearer $DRI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "clinical trial immunotherapy data",
    "sources": ["pubmed", "openalex"],
    "options": {
      "timeoutMs": 1500
    }
  }'

Unified Connector Interface

All integrations implement a standardized interface to avoid connector debt. Read our complete guide on eliminating the connector tax to review the TypeScript blueprint and interface definitions.

Handling Query Latency

Parallel queries to third-party government databases are subject to network stalls. To protect your application from latency spikes, implement circuit breakers and timeout fallbacks as detailed in our guide on resilient federated search.