How driftloom/vaeloom works
Vaeloom is a memory-first career and productivity platform for students and early-career professionals. Its main workflow is to collect user-created files and communications, parse and organize them into searchable memory, then use scoped agents to surface useful information and support career tasks such as resume tailoring and job applications. The API also provides orchestration, model access, and enterprise controls. Some major capabilities are documented in the README but their runtime wiring was not sampled, so those links are shown as unverified rather than asserted as direct calls.
Main entry points: pipeline.py Document Ingestion · main.py API Backend · engine.py Request Routing · search_service.py Hybrid Search
Components 21
User Experience
- Web App
apps/web/src/app - API Backend
apps/api/src/api/main.py
Knowledge Pipeline
- Document Ingestion
apps/api/src/api/ingestion/pipeline.py - Content Parsers
apps/api/src/api/ingestion/parsers.py - Chunking
apps/api/src/api/ingestion/chunking.py - Deduplication
apps/api/src/api/ingestion/dedup.py - Memory Service
apps/api/src/api/services/memory_service.py - Knowledge Graph
apps/api/src/api/services/knowledge_graph_service.py - Hybrid Search
apps/api/src/api/services/search_service.py
Intelligence
- Request Routing
apps/api/src/api/orchestrator/routing/engine.py - Agent Capabilities
apps/api/src/api/orchestrator/capability_registry.py - Agent Policy
packages/agent-policy/src/vaeloom_agent_policy/engine.py - Agent Runtime
apps/api/src/api/services/agent_runtime.py - Model Service
apps/api/src/api/services/llm_service.py
Career Workflows
- Resume Service
apps/api/src/api/services/resume_service.py - Applications
apps/api/src/api/services/application_service.py - Opportunity Matching
apps/api/src/api/services/opportunity_matcher.py
Platform Services
- Relational Store
apps/api/src/api/database.py - Domain Records
apps/api/src/api/models/schema.py - Tenant Controls
apps/api/src/api/middleware/tenant.py
Other components
- User
Connections 24
- User → Web App: uses
README.md - Web App → API Backend: requests
- API Backend → Document Ingestion: accepts content
- Document Ingestion → Content Parsers: parses
apps/api/src/api/ingestion/pipeline.py - Document Ingestion → Chunking: chunks
apps/api/src/api/ingestion/pipeline.py - Document Ingestion → Deduplication: deduplicates
apps/api/src/api/ingestion/pipeline.py - Document Ingestion → Relational Store: persists
apps/api/src/api/ingestion/pipeline.py - Document Ingestion → Domain Records: creates records
apps/api/src/api/ingestion/pipeline.py - Document Ingestion → Model Service: embeds content
apps/api/src/api/ingestion/pipeline.py - Document Ingestion → Knowledge Graph: mirrors entities
apps/api/src/api/ingestion/pipeline.py - Document Ingestion → Memory Service: stores extracted facts
apps/api/src/api/ingestion/pipeline.py - Memory Service → Hybrid Search: supports retrieval
- Knowledge Graph → Hybrid Search: supports traversal
README.md - API Backend → Hybrid Search: searches workspace
README.md - API Backend → Request Routing: routes requests
- Request Routing → Agent Capabilities: plans capabilities
README.md - Request Routing → Agent Runtime: plans execution
README.md - Request Routing → Agent Policy: enforces scope
README.md - Agent Runtime → Model Service: uses models
README.md - API Backend → Resume Service: supports resume tasks
README.md - API Backend → Applications: tracks applications
README.md - API Backend → Opportunity Matching: matches opportunities
README.md - API Backend → Tenant Controls: isolates tenants
README.md - Relational Store → Domain Records: stores records
apps/api/src/api/database.py