AI Backend Framework
Summary
AF Deep Research is an autonomous AI backend executing large-scale, self-correcting research workflows via API or CLI. It orchestrates thousands of parallel agents to search, filter, and iterate until quality thresholds are met, returning structured JSON with entities and evidence.
Highlights
- Orchestrates approximately 10,000 logical agent invocations per research task in parallel streams.
- Features self-correcting research loops that generate new sub-queries when information gaps are detected.
- Provides a simple developer experience via CLI or REST API with SSE streaming for real-time progress updates.
- Returns structured data including typed entities, relationships, evidence, and full cited documents.
- Supports self-hosting via Docker or direct Python execution with configurable workspaces.
auto-generated
Context
Audience
AI developers and researchers building autonomous research applications
DomainArtificial Intelligence
Formatopen source software
Accessopen source
Topics
Related
Autonomous AgentsREST APIsResearch AutomationDeep ResearchAgentField RuntimeOpenRouter API