API Docs

Configurable RAG and Agentic Platform Guides

This page covers architecture, governance, and strategy guidance. DAILOGUE supports a canonical 5-level scope hierarchy (platform, application, customer, organization, project) with WHO/TARGET header-based authorization. Use API Reference for endpoint contracts.

Platform-first reference for technical decision makersPublic-safe endpoint subset on platform.dailogue.aiInternal full API docs remain on api.dailogue.ai

Start Quickstart Open API Reference View Configuration Matrix Internal API Explorer (SSO)

At a glance

Reading order: 1) Quickstart, 2) capability and ecosystem choices, 3) governance and evaluation controls, 4) move to reference for request/response contracts.

Quickstart

Start from platform controls, then move into curated reference details.

1. Authenticate

Use tenant-scoped credentials and role policies to access approved API surfaces.

2. Configure a workload

Select model routing, RAG profile, and agent policies by tenant and application context.

3. Observe and improve

Track groundedness, relevance, and latency signals to continuously optimize outcomes.

Platform Capabilities

Built for product, engineering, and AI teams adopting enterprise GenAI systems with configurable control surfaces.

RAG Orchestration

Configure retrieval quality without replacing your platform stack.

• Hybrid retrieval, reranking, and context assembly controls
• Grounded generation strategies for regulated workflows
• Policy-driven quality controls across tenants and applications

Agentic Execution

Run multi-step automation with approvals, constraints, and audit trails.

• Workflow graphs, tool permissions, and checkpoint policies
• Human-in-the-loop review paths for critical business outputs
• Future-ready path for advanced tool and computer-use workflows

Evaluation and Observability

Measure quality, groundedness, and performance with operational visibility.

• Evaluation surfaces for relevance, grounding, and response quality
• Traceable pipeline execution across ingestion, retrieval, and generation
• Tenant-aware telemetry for compliance and operational governance

Configurable AI Stack

The platform supports configurable providers and libraries through policy controls, not ecosystem rewrites.

Provider and library configuration matrix
Layer	Configurable Choices	Example Providers/Libraries	Selection Policy
Model Routing	Provider routing, model family, fallback policy, guardrails	OpenAI, Anthropic Claude, Google Gemini, Mistral	Optimize by cost, quality, and latency profile per workload
Embeddings and Reranking	Embedding model, reranker strategy, cross-encoder stage	Voyage rerankers, provider-native rerank, policy-selected rerank pipelines	Raise retrieval precision for high-risk or regulated outputs
Retrieval Stack	Semantic, lexical, hybrid retrieval, contextual compression	Hybrid vector + keyword retrieval with policy-driven assembly	Balance recall and precision by tenant/application policy
Agent Orchestration	Workflow graph, tool permissions, approval checkpoints	LangGraph-style orchestration and internal workflow templates	Automate with human oversight for critical business flows

RAG Strategy Patterns

Configure retrieval and grounding policies by vertical, tenant, and application requirements.

Semantic Retrieval

Vector-first retrieval for concept-level matches across long enterprise content.

When to use: Use for knowledge-heavy workloads where meaning is more important than exact keyword match.

Tradeoff: Higher recall, may need reranking to improve precision.

Hybrid Retrieval

Combine semantic retrieval with lexical search for stronger recall and precision.

When to use: Use for legal, compliance, and reporting flows that require both intent and exact term coverage.

Tradeoff: More configuration and runtime complexity, better groundedness outcomes.

Reranking and Cross-Encoder Stage

Apply a second-stage relevance ranker to tighten top-k evidence quality.

When to use: Use for customer-facing and compliance-critical outputs where precision matters most.

Tradeoff: Adds latency/cost, significantly improves relevance quality.

Query Understanding

Rewrite, decompose, and classify queries before retrieval execution.

When to use: Use for ambiguous, multi-part, or domain-specific user questions.

Tradeoff: Extra orchestration step, improves retrieval targeting.

Groundedness Controls

Attach citation and grounding policies to generation and approval flows.

When to use: Use when outputs are compliance-sensitive, auditable, or externally published.

Tradeoff: Can reduce generation flexibility, improves trust and auditability.

Configuration Scope by Governance Layer

Controls inherit across the canonical hierarchy: platform → application → customer → organization → project. WHO headers identify the actor and TARGET headers identify the operated scope.

Configuration scope map
Scope	What You Configure	Operational Outcome
Platform	Global model policy, safety baselines, observability defaults, and audit controls	Consistent governance guardrails for every downstream tenant scope
Application	Application-level routing policy, prompt bundles, and tool boundary defaults	Shared vertical controls without duplicating service code
Customer	Customer policy overlays for RAG thresholds, quality targets, and data boundaries	Per-customer control while preserving application defaults
Organization	Organization-level workflow approvals, review policy, and publication guardrails	Department-level governance for regulated output paths
Project	Project runtime tuning for prompts, retrieval strategy, and execution checkpoints	Fine-grained optimization for delivery teams and workloads

Platform

Set default model safety baseline and required groundedness checks.

Consistent compliance guardrails across all tenants and applications.

Application + Customer

Switch retrieval profile to hybrid + reranking for legal tenants.

Domain-specific quality gains without branching platform code.

Organization + Project

Tune query understanding and agent approvals for newsletter publication by team.

Faster iteration per product workflow with governed automation.

Evaluation and Observability Controls

Enterprise adoption requires grounded outputs, measurable quality, and operational traceability.

Groundedness Scorecards

Track groundedness and citation quality across workloads and release cycles.

Retrieval Quality Monitoring

Measure relevance, recall, and rerank lift to catch drift before it impacts users.

End-to-End Execution Traces

Trace ingest, process, retrieve, and generate stages for root-cause analysis.

Policy and Approval Audit Trail

Maintain tenant-aware logs for workflow approvals and regulated outputs.

Grounded response strategy with retrieval-aware context assemblyEvaluation signals for quality, relevance, and policy alignmentObservability across pipeline execution stages and model routingMulti-tenant isolation and role-based access boundaries

Public API Coverage Snapshot

Curated operation groups published for external-safe reference. Full endpoint surface remains internal.

API reference operation groups and status
Group	Operations	Status
Platform Health	1	Internal GA
Ingestion	3	Internal GA
Processing	1	Internal GA
LLM	1	Internal GA
Storage	1	Internal GA
Retrieval (RAG)	1	Public Preview
Agentic	1	Public Preview

Code Examples

Common integration patterns using the DAILOGUE API with WHO/TARGET authorization headers.

Authentication with WHO/TARGET Headers

curl -X POST https://api.dailogue.ai/api/v2/retrieval/query \
  -H "Authorization: Bearer <token>" \
  -H "X-Application-Tenant-Id: <app-tenant-id>" \
  -H "X-Application-Id: <app-entity-id>" \
  -H "X-Target-Tenant-Id: <target-tenant-id>" \
  -H "X-Target-Tenant-Scope: customer" \
  -H "Content-Type: application/json"

Retrieval Query with Hybrid Strategy

{
  "query": "What are the ESG reporting requirements for Q4?",
  "strategy": "hybrid_rerank",
  "topK": 8
}

Combines vector + keyword retrieval with cross-encoder reranking for grounded context assembly.

Create Ingestion Source

curl -X POST https://api.dailogue.ai/api/v2/ingestion/sources \
  -H "Authorization: Bearer <token>" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "Weekly Newsletter Feed",
    "sourceType": "rss",
    "config": {
      "url": "https://example.com/feed.xml",
      "schedule": "0 6 * * 1"
    }
  }'

API Reference Entry Point

Use API Reference for endpoint-level request/response schemas and interactive exploration.

Separate Reference Experience

Guides explain architecture, configuration, and operational patterns. Reference focuses on endpoint contracts.

Go to API Reference Internal API Explorer (SSO)