How binarystudioacademy/bsa-2026-promtomat works
Promptomat is an intelligent prompt-search, evaluation, and autocomplete system for AI coding tools. Its product workflow centers on users organizing workspaces and prompts, searching prompt material semantically, and generating composed prompts; the backend also supports analytics, authentication, and repository bindings. The README documents a local MCP server that lets coding agents access the backend with an API token. The sampled code confirms prompt API delegation and embedding generation, persistence, and nearest-neighbor queries. Other feature links reflect documented capabilities and module boundaries; their internal call wiring was not sampled.
Main entry points: index.ts Backend API · prompt-embedding.service.ts Embedding Service · analytics Prompt Analytics · prompt.controller.ts Prompt API
Components 21
Product API
- Backend API
apps/backend/src/index.ts - Prompt Analytics
apps/backend/src/modules/analytics
Prompt Search
- Prompt API
apps/backend/src/modules/prompts/prompt.controller.ts - Prompt Operations
apps/backend/src/modules/prompts - Embedding Service
apps/backend/src/modules/prompt-embeddings/prompt-embedding.service.ts - Vector Queries
apps/backend/src/modules/prompt-embeddings/prompt-embedding.repository.ts - Prompt Composition
apps/backend/src/modules/composed-prompts/composed-prompt.service.ts
Workspace Data
- Workspace Management
apps/backend/src/modules/workspaces/workspace.controller.ts - Contributors
apps/backend/src/modules/contributors/contributor.controller.ts - Labels
apps/backend/src/modules/labels - Repository Bindings
apps/backend/src/modules/repository-bindings/repository-binding.controller.ts
Platform Services
- Authentication
apps/backend/src/modules/auth - API Tokens
apps/backend/src/modules/api-tokens - Embedding Model
apps/backend/src/libs/modules/embedding/embedding.ts - Text Generation
apps/backend/src/libs/modules/generator/generator.ts
Agent Access
- Local MCP Server
apps/mcp
Other components
- Users
- Coding Agents
- Frontend
- Bedrock
- PostgreSQL pgvector
Connections 25
- Users → Frontend: uses
- Frontend → Backend API: calls API
- Coding Agents → Local MCP Server: starts locally
- Local MCP Server → Backend API: calls API
- Backend API → Authentication: authenticates
- Backend API → API Tokens: validates tokens
- Backend API → Prompt API: dispatches routes
- Prompt API → Prompt Operations: delegates
- Prompt Operations → Embedding Service: searches semantically
- Embedding Service → Embedding Model: embeds text
- Embedding Service → Vector Queries: queries and stores
- Vector Queries → PostgreSQL pgvector: reads and writes
- Backend API → Prompt Composition: dispatches routes
- Prompt Composition → Text Generation: generates output
- Text Generation → Bedrock: invokes model
- Backend API → Workspace Management: dispatches routes
- Backend API → Contributors: dispatches routes
- Backend API → Labels: dispatches routes
- Backend API → Repository Bindings: dispatches routes
- Backend API → Prompt Analytics: dispatches routes
- Workspace Management → PostgreSQL pgvector: persists workspace data
- Contributors → PostgreSQL pgvector: persists membership
- Labels → PostgreSQL pgvector: persists labels
- Repository Bindings → PostgreSQL pgvector: persists bindings
- Prompt Analytics → PostgreSQL pgvector: queries metrics