Skip to content

Sri-Krishna-V/Recall

BranchesTags
 
 

Repository files navigation

Recall

The Self-Learning MCP Data Agent

From natural language queries to SQL insights with 6 layers of context and automatic learning.

Version License

Python FastAPI MCP-Enabled Docker

Ask DeepWiki

About · Architecture · 6 Layers · Getting Started · Tech Stack · Demo

Table of Contents

Recall

About the Project

Recall is a production-ready Self-Learning MCP Data Agent that delivers insights, not just SQL results. Inspired by OpenAI's internal data agent, Recall grounds every query in 6 layers of context and improves automatically with every interaction.

Traditional SQL agents fail repeatedly on the same errors. They return raw SQL output that users still need to interpret. They lack observability into their reasoning process.

Recall is different. It treats errors as learning opportunities, synthesizes natural language insights from query results, and exposes every step of its pipeline for full transparency.

Key Transformations

flowchart LR
    subgraph before [Traditional]
        SQL[Raw SQL Results]
        Static[Static Knowledge]
        Opaque[Opaque Process]
        Error[Error-Prone]
    end
    subgraph after [Recall]
        Insights[Natural Language Insights]
        SelfLearn[Curated + Discovered]
        Observable[7-Step Visualization]
        SelfCorrect[Self-Correcting]
    end
    SQL -->|"SQL to Insights"| Insights
    Static -->|"Static to Self-Learning"| SelfLearn
    Opaque -->|"Opaque to Observable"| Observable
    Error -->|"Error-Prone to Self-Correcting"| SelfCorrect
    
    style SQL fill:#546e7a,stroke:#90a4ae,stroke-width:2px,color:#eceff1
    style Static fill:#546e7a,stroke:#90a4ae,stroke-width:2px,color:#eceff1
    style Opaque fill:#546e7a,stroke:#90a4ae,stroke-width:2px,color:#eceff1
    style Error fill:#546e7a,stroke:#90a4ae,stroke-width:2px,color:#eceff1
    style Insights fill:#1b5e20,stroke:#66bb6a,stroke-width:2px,color:#f1f8e9
    style SelfLearn fill:#1b5e20,stroke:#66bb6a,stroke-width:2px,color:#f1f8e9
    style Observable fill:#1b5e20,stroke:#66bb6a,stroke-width:2px,color:#f1f8e9
    style SelfCorrect fill:#1b5e20,stroke:#66bb6a,stroke-width:2px,color:#f1f8e9
Loading

Built for Hackathons, Ready for Production. Recall was built for the "2 Fast 2 MCP" hackathon with OpenTelemetry traces, Prometheus metrics, Docker + Kubernetes deployment, and a ChatGPT-style UI with process visualization.

System Architecture

Recall follows a context-grounded agent architecture where every SQL query is enriched with 6 layers of contextual knowledge before execution.

Recall Architecture

Detailed Component Flow

graph TB
    User[User Query] -->|Natural Language| UI[React Frontend]
    UI -->|POST Request| API[FastAPI Server]
    
    API -->|MCP Protocol| Agent[Recall Agent]
    
    Agent -->|Step 1| Parse[Parse Query]
    Parse -->|Step 2| Knowledge[Search Knowledge Base]
    Knowledge -->|Step 3| Learnings[Retrieve Learnings]
    Learnings -->|Step 4| Schema[Introspect Schema]
    Schema -->|Step 5| SQLGen[Generate SQL]
    SQLGen -->|Step 6| Execute[Execute Query]
    Execute -->|Step 7| Format[Format Insights]
    
    Knowledge -->|Retrieve| VectorDB[(pgvector)]
    Learnings -->|Retrieve| LearningDB[(Learning Machine)]
    Execute -->|Query| PostgreSQL[(PostgreSQL)]
    
    SQLGen -->|LLM Call| Gemini[Google Gemini]
    Format -->|LLM Call| Gemini
    
    Format -->|Natural Language Answer| UI
    UI -->|Display with Process Viz| User
    
    API -->|Traces| OTEL[OpenTelemetry]
    OTEL -->|Metrics| Prometheus[Prometheus]
    
    style User fill:#546e7a,stroke:#90a4ae,stroke-width:2px,color:#eceff1
    style UI fill:#1565c0,stroke:#4a9eff,stroke-width:2px,color:#e3f2fd
    style API fill:#0d47a1,stroke:#42a5f5,stroke-width:2px,color:#e3f2fd
    style Agent fill:#1b5e20,stroke:#66bb6a,stroke-width:3px,color:#f1f8e9
    style Parse fill:#1a3d2b,stroke:#4caf7d,stroke-width:2px,color:#e8f5e9
    style Knowledge fill:#1a3d2b,stroke:#4caf7d,stroke-width:2px,color:#e8f5e9
    style Learnings fill:#1a3d2b,stroke:#4caf7d,stroke-width:2px,color:#e8f5e9
    style Schema fill:#1a3d2b,stroke:#4caf7d,stroke-width:2px,color:#e8f5e9
    style SQLGen fill:#1a3d2b,stroke:#4caf7d,stroke-width:2px,color:#e8f5e9
    style Execute fill:#1a3d2b,stroke:#4caf7d,stroke-width:2px,color:#e8f5e9
    style Format fill:#1a3d2b,stroke:#4caf7d,stroke-width:2px,color:#e8f5e9
    style VectorDB fill:#006064,stroke:#00e5ff,stroke-width:2px,color:#e0f7fa
    style LearningDB fill:#006064,stroke:#00e5ff,stroke-width:2px,color:#e0f7fa
    style PostgreSQL fill:#bf360c,stroke:#ff8a65,stroke-width:2px,color:#fbe9e7
    style Gemini fill:#4a148c,stroke:#ce93d8,stroke-width:2px,color:#f3e5f5
    style OTEL fill:#004d40,stroke:#4db6ac,stroke-width:2px,color:#e0f2f1
    style Prometheus fill:#004d40,stroke:#4db6ac,stroke-width:2px,color:#e0f2f1
Loading

The 6 Layers of Context

Most SQL agents use 1-2 layers of context. Recall uses 6 distinct layers that are semantically retrieved and combined at query time.

flowchart TB
    L1[Layer 1: Table Usage Patterns<br/>tables/*.json]
    L2[Layer 2: Business Rules & Metrics<br/>business/*.json]
    L3[Layer 3: Query Patterns<br/>queries/*.sql]
    L4[Layer 4: Institutional Knowledge<br/>Exa MCP]
    L5[Layer 5: Dynamic Learnings<br/>Learning Machine]
    L6[Layer 6: Runtime Context<br/>introspect_schema]
    
    L1 --> L2 --> L3 --> L4 --> L5 --> L6
    
    style L1 fill:#1565c0,stroke:#4a9eff,stroke-width:2px,color:#e3f2fd
    style L2 fill:#4a148c,stroke:#ce93d8,stroke-width:2px,color:#f3e5f5
    style L3 fill:#1b5e20,stroke:#66bb6a,stroke-width:2px,color:#f1f8e9
    style L4 fill:#bf360c,stroke:#ff8a65,stroke-width:2px,color:#fbe9e7
    style L5 fill:#6a1b9a,stroke:#ba68c8,stroke-width:2px,color:#f3e5f5
    style L6 fill:#006064,stroke:#00e5ff,stroke-width:2px,color:#e0f7fa
Loading

Example (Layer 1):

{
  "table": "race_wins",
  "columns": ["driver_name", "race_name", "date"],
  "common_joins": ["drivers_championship"],
  "date_format": "DD Mon YYYY"
}

Demo Video

Watch the Recall demo video

Click the thumbnail above to watch the full project demo on YouTube.

ChatGPT-Style Interface with Process Visualization

flowchart LR
    P1[1. Parse] --> P2[2. Knowledge]
    P2 --> P3[3. Learnings]
    P3 --> P4[4. Schema]
    P4 --> P5[5. SQL]
    P5 --> P6[6. Execute]
    P6 --> P7[7. Format]
    
    style P1 fill:#1565c0,stroke:#4a9eff,stroke-width:2px,color:#e3f2fd
    style P2 fill:#4a148c,stroke:#ce93d8,stroke-width:2px,color:#f3e5f5
    style P3 fill:#1b5e20,stroke:#66bb6a,stroke-width:2px,color:#f1f8e9
    style P4 fill:#bf360c,stroke:#ff8a65,stroke-width:2px,color:#fbe9e7
    style P5 fill:#6a1b9a,stroke:#ba68c8,stroke-width:2px,color:#f3e5f5
    style P6 fill:#006064,stroke:#00e5ff,stroke-width:2px,color:#e0f7fa
    style P7 fill:#004d40,stroke:#4db6ac,stroke-width:2px,color:#e0f2f1
Loading

ChatGPT-style UI with dark theme, expandable data views per step, natural language responses, and copy-to-clipboard for SQL.

Tech Stack

flowchart TB
    subgraph core [Core Intelligence]
        Agno[Agno MCP]
        MCP[MCP Protocol]
        Gemini[Gemini]
        Ollama[Ollama / nomic-embed]
    end
    
    subgraph backend [Backend & API]
        FastAPI[FastAPI]
        Python[Python 3.12]
        Pydantic[Pydantic v2]
        PG[(PostgreSQL + pgvector)]
    end
    
    subgraph knowledge [Knowledge]
        pgvec[pgvector]
        JSON[JSON Knowledge]
        Learning[Learning Machine]
    end
    
    subgraph frontend [Frontend]
        React[React 18 + Vite]
        Tailwind[Tailwind CSS]
        Inter[Inter Font]
    end
    
    subgraph infra [Infrastructure]
        Docker[Docker + K8s]
        OTEL[OpenTelemetry]
        Prom[Prometheus + Grafana]
    end
    
    style Agno fill:#1b5e20,stroke:#66bb6a,stroke-width:2px,color:#f1f8e9
    style MCP fill:#1b5e20,stroke:#66bb6a,stroke-width:2px,color:#f1f8e9
    style Gemini fill:#4a148c,stroke:#ce93d8,stroke-width:2px,color:#f3e5f5
    style Ollama fill:#4a148c,stroke:#ce93d8,stroke-width:2px,color:#f3e5f5
    style FastAPI fill:#1a3d2b,stroke:#4caf7d,stroke-width:2px,color:#e8f5e9
    style Python fill:#1a3d2b,stroke:#4caf7d,stroke-width:2px,color:#e8f5e9
    style Pydantic fill:#1a3d2b,stroke:#4caf7d,stroke-width:2px,color:#e8f5e9
    style PG fill:#006064,stroke:#00e5ff,stroke-width:2px,color:#e0f7fa
    style pgvec fill:#006064,stroke:#00e5ff,stroke-width:2px,color:#e0f7fa
    style JSON fill:#bf360c,stroke:#ff8a65,stroke-width:2px,color:#fbe9e7
    style Learning fill:#006064,stroke:#00e5ff,stroke-width:2px,color:#e0f7fa
    style React fill:#1565c0,stroke:#4a9eff,stroke-width:2px,color:#e3f2fd
    style Tailwind fill:#1565c0,stroke:#4a9eff,stroke-width:2px,color:#e3f2fd
    style Inter fill:#1565c0,stroke:#4a9eff,stroke-width:2px,color:#e3f2fd
    style Docker fill:#004d40,stroke:#4db6ac,stroke-width:2px,color:#e0f2f1
    style OTEL fill:#004d40,stroke:#4db6ac,stroke-width:2px,color:#e0f2f1
    style Prom fill:#004d40,stroke:#4db6ac,stroke-width:2px,color:#e0f2f1
Loading

Getting Started

Prerequisites

  • Docker & Docker Compose (for simplest setup)
  • Gemini API Key from Google AI Studio
  • Ollama installed on host machine for embeddings:
    • nomic-embed-text:latest (embeddings only)
  • Node.js 18+ (for frontend development)
  • Python 3.12+ (for backend development)

Quick Start (Docker)

flowchart TB
    S1[1. Clone] --> S2[2. Config]
    S2 --> S3[3. Ollama]
    S3 --> S4[4. Docker Up]
    S4 --> S5[5. Load Data]
    S5 --> S6[6. Load Knowledge]
    S6 --> S7[7. Frontend]
    S7 --> S8[8. Open Browser]
    
    style S1 fill:#1565c0,stroke:#4a9eff,stroke-width:2px,color:#e3f2fd
    style S2 fill:#4a148c,stroke:#ce93d8,stroke-width:2px,color:#f3e5f5
    style S3 fill:#1b5e20,stroke:#66bb6a,stroke-width:2px,color:#f1f8e9
    style S4 fill:#bf360c,stroke:#ff8a65,stroke-width:2px,color:#fbe9e7
    style S5 fill:#6a1b9a,stroke:#ba68c8,stroke-width:2px,color:#f3e5f5
    style S6 fill:#006064,stroke:#00e5ff,stroke-width:2px,color:#e0f7fa
    style S7 fill:#1565c0,stroke:#4a9eff,stroke-width:2px,color:#e3f2fd
    style S8 fill:#1b5e20,stroke:#66bb6a,stroke-width:2px,color:#f1f8e9
Loading 
git clone https://github.com/Keerthivasan-Venkitajalam/Recall.git && cd Recall
cp example.env .env   # Add GEMINI_API_KEY
ollama pull nomic-embed-text:latest
docker compose up -d --build
docker exec recall-api python -m recall.scripts.load_data
docker exec recall-api python -m recall.scripts.load_knowledge
cd web && npm install && npm run dev
# Open http://localhost:5174/ and http://localhost:8000/docs

Verify Health

curl http://localhost:8000/health/dependencies | jq

Expected response:

{
  "status": "healthy",
  "checks": {
    "database": { "status": "healthy" },
    "vector_db_knowledge": { "status": "healthy", "sample_count": 8 },
    "vector_db_learnings": { "status": "healthy" }
  }
}

Configuration

Recall is configured via environment variables in .env:

Variable Description Default
GEMINI_API_KEY Google Gemini API key (required)
MODEL_PROVIDER LLM provider (not used with Gemini) -
EMBEDDER_PROVIDER Embedding provider (nomic via Ollama) nomic
OLLAMA_BASE_URL Ollama API endpoint for embeddings http://host.docker.internal:11434
DB_HOST PostgreSQL host recall-db
DB_PORT PostgreSQL port 5432
DB_NAME Database name ai
DB_USER Database user keerthi
DB_PASSWORD Database password *****
LOG_LEVEL Logging level INFO

Docker Compose Configuration

The compose.yaml includes:

  • recall-api: FastAPI backend with Agno MCP framework and Gemini
  • recall-db: PostgreSQL 18 with pgvector extension
  • Health checks on both services
  • Persistent volumes for database
  • Host network access via extra_hosts for Ollama embeddings

How the Dual Learning System Works

flowchart TB
    User[User Query] --> Agent[Recall Agent]
    Agent --> Static[Static Knowledge<br/>recall/knowledge/]
    Agent --> Dynamic[Dynamic Learnings<br/>Learning Machine]
    Static --> VectorDB[(pgvector)]
    Dynamic --> VectorDB
    VectorDB --> Future[Future Queries]
    
    subgraph curated [Curated by You]
        Static
    end
    subgraph discovered [Auto-Discovered]
        Dynamic
    end
    
    style User fill:#546e7a,stroke:#90a4ae,stroke-width:2px,color:#eceff1
    style Agent fill:#1b5e20,stroke:#66bb6a,stroke-width:2px,color:#f1f8e9
    style Static fill:#1565c0,stroke:#4a9eff,stroke-width:2px,color:#e3f2fd
    style Dynamic fill:#4a148c,stroke:#ce93d8,stroke-width:2px,color:#f3e5f5
    style VectorDB fill:#006064,stroke:#00e5ff,stroke-width:2px,color:#e0f7fa
    style Future fill:#1b5e20,stroke:#66bb6a,stroke-width:2px,color:#f1f8e9
Loading

Example learning (auto-captured):

{
  "pattern": "date_parsing_f1",
  "context": "F1 race_wins table has dates as 'DD Mon YYYY' text format",
  "solution": "Use TO_DATE(date, 'DD Mon YYYY') for year extraction"
}

Development

Project Structure

recall/
├── agents.py                  # Recall and Reasoning agents
├── paths.py                   # Path constants
├── context/                   # Context layer implementations
│   ├── semantic_model.py     # Layer 1: Table usage
│   └── business_rules.py     # Layer 2: Business rules
├── tools/                     # Agent tools
│   ├── introspect.py         # Layer 6: Runtime context
│   └── save_query.py         # Save validated queries
├── scripts/
│   ├── load_data.py          # Load F1 sample data
│   └── load_knowledge.py     # Load knowledge files
├── knowledge/                 # Static knowledge files
│   ├── tables/               # Layer 1 schemas
│   ├── queries/              # Layer 3 patterns
│   └── business/             # Layer 2 rules
├── evals/
│   ├── test_cases.py         # Test cases with golden SQL
│   ├── grader.py             # LLM-based response grader
│   └── run_evals.py          # Run evaluations

web/
├── src/
│   ├── App.jsx               # Main React component
│   ├── main.jsx              # React entry point
│   └── styles.css            # Tailwind + custom styles
├── index.html                # HTML template
├── vite.config.js            # Vite configuration
├── tailwind.config.cjs       # Tailwind configuration
└── package.json              # Frontend dependencies

db/
├── session.py                # PostgreSQL session factory
└── url.py                    # Database URL builder

k8s/
└── base/                     # Kubernetes manifests

Running Tests

# Backend tests
python -m recall.evals.run_evals              # All evals (string matching)
python -m recall.evals.run_evals -c basic     # Specific category
python -m recall.evals.run_evals -v           # Verbose mode
python -m recall.evals.run_evals -g           # Use LLM grader
python -m recall.evals.run_evals -r           # Compare against golden SQL results
python -m recall.evals.run_evals -g -r -v     # All modes combined

# Frontend (in web/ directory)
npm test                                     # Run tests
npm run build                                # Production build

Code Quality

# Backend
./scripts/format.sh      # Format code (black, isort)
./scripts/validate.sh    # Lint + type check (ruff, mypy)

# Frontend
cd web
npm run lint             # ESLint

Deployment

Docker (Production)

# Build and deploy
docker compose -f compose.yaml up -d --build

# View logs
docker logs -f recall-api
docker logs -f recall-db

# Health check
curl http://localhost:8000/health/dependencies

Kubernetes

# Apply configurations
kubectl apply -k k8s/base/

# Check status
kubectl get pods -n recall
kubectl logs -n recall deployment/recall-api

# Port forward for local access
kubectl port-forward -n recall svc/recall-api 8000:8000

Advanced Use Cases

1. Complex Data Analysis — "Who won the most races in 2019?" 
flowchart LR
    Parse[Parse] --> Know[Knowledge]
    Know --> Learn[Learnings]
    Learn --> Schema[Schema]
    Schema --> SQL[SQL Gen]
    SQL --> Exec[Execute]
    Exec --> Format[Format]
    Format --> Result["Lewis Hamilton, 11 wins"]
    
    style Parse fill:#1565c0,stroke:#4a9eff,color:#e3f2fd
    style Know fill:#4a148c,stroke:#ce93d8,color:#f3e5f5
    style Learn fill:#1b5e20,stroke:#66bb6a,color:#f1f8e9
    style Schema fill:#bf360c,stroke:#ff8a65,color:#fbe9e7
    style SQL fill:#6a1b9a,stroke:#ba68c8,color:#f3e5f5
    style Exec fill:#006064,stroke:#00e5ff,color:#e0f7fa
    style Format fill:#004d40,stroke:#4db6ac,color:#e0f2f1
    style Result fill:#1b5e20,stroke:#66bb6a,color:#f1f8e9
Loading
2. Self-Correction on Error — "Constructor standings for 2020" 
flowchart LR
    Try1[year = 2020] --> Err[Error]
    Err --> Capture[Learning: use season]
    Capture --> Retry[season = 2020]
    Retry --> Success[Success]
    Success --> Save[Saved for future]
    
    style Try1 fill:#546e7a,stroke:#90a4ae,color:#eceff1
    style Err fill:#b71c1c,stroke:#ef5350,color:#ffebee
    style Capture fill:#bf360c,stroke:#ff8a65,color:#fbe9e7
    style Retry fill:#1565c0,stroke:#4a9eff,color:#e3f2fd
    style Success fill:#1b5e20,stroke:#66bb6a,color:#f1f8e9
    style Save fill:#1b5e20,stroke:#66bb6a,color:#f1f8e9
Loading
3. Cross-Table Insights — "Fastest lap times in Monaco 2019" 
flowchart TB
    L1[Layer 1: fastest_laps + race_results]
    L2[Layer 2: Monaco = race_name]
    L3[Layer 3: Query pattern]
    L4[Layer 4: Web context]
    L1 --> Join[Complex Join]
    L2 --> Join
    L3 --> Join
    L4 --> Join
    Join --> FirstTry[First-try success]
    
    style L1 fill:#1565c0,stroke:#4a9eff,color:#e3f2fd
    style L2 fill:#4a148c,stroke:#ce93d8,color:#f3e5f5
    style L3 fill:#1b5e20,stroke:#66bb6a,color:#f1f8e9
    style L4 fill:#bf360c,stroke:#ff8a65,color:#fbe9e7
    style Join fill:#6a1b9a,stroke:#ba68c8,color:#f3e5f5
    style FirstTry fill:#1b5e20,stroke:#66bb6a,color:#f1f8e9
Loading

Observability & Monitoring

flowchart LR
    Req[Request] --> Parse[parse_query]
    Parse --> Know[search_knowledge]
    Know --> Learn[search_learnings]
    Learn --> Schema[introspect_schema]
    Schema --> SQL[generate_sql]
    SQL --> Exec[execute_query]
    Exec --> Format[format_insight]
    Format --> OTEL[OpenTelemetry]
    OTEL --> Prom[Prometheus]
    Prom --> Grafana[Grafana]
    
    style Req fill:#546e7a,stroke:#90a4ae,color:#eceff1
    style Parse fill:#1a3d2b,stroke:#4caf7d,color:#e8f5e9
    style Know fill:#1a3d2b,stroke:#4caf7d,color:#e8f5e9
    style Learn fill:#1a3d2b,stroke:#4caf7d,color:#e8f5e9
    style Schema fill:#1a3d2b,stroke:#4caf7d,color:#e8f5e9
    style SQL fill:#1a3d2b,stroke:#4caf7d,color:#e8f5e9
    style Exec fill:#1a3d2b,stroke:#4caf7d,color:#e8f5e9
    style Format fill:#1a3d2b,stroke:#4caf7d,color:#e8f5e9
    style OTEL fill:#004d40,stroke:#4db6ac,color:#e0f2f1
    style Prom fill:#bf360c,stroke:#ff8a65,color:#fbe9e7
    style Grafana fill:#4a148c,stroke:#ce93d8,color:#f3e5f5
Loading
Prometheus Metrics (at /metrics) 
recall_queries_total{status="success"} 127
recall_query_duration_seconds_bucket{le="5.0"} 120
recall_knowledge_searches_total{layer="table_usage"} 127
recall_llm_tokens_total{model="gemini"} 52341

Grafana dashboard: query throughput, latency percentiles (p50/p95/p99), token usage, learning rate.

Troubleshooting

Common issues and quick fixes. Ensure GEMINI_API_KEY, Ollama (nomic-embed-text), and Docker are configured.

Issue Solution
"Gemini API error" Verify GEMINI_API_KEY in .env is valid. Check API quota at Google AI Studio.
"Knowledge base empty" Run docker exec recall-api python -m recall.scripts.load_knowledge to index documents.
"Frontend can't connect to API" Check VITE_API_URL in web/.env. Ensure backend is running on port 8000.
"Embedding model error" Ensure Ollama is running and nomic-embed-text is pulled: ollama list.
"Database connection failed" Check DB_* environment variables. Ensure recall-db container is healthy: docker ps.
"Process monitor not expanding" Click on completed steps (green checkmark). Only completed steps have data.

Contributing

Contributions are welcome! Whether you're fixing bugs, adding new context layers, improving the UI, or enhancing documentation, your help is appreciated.

  1. Fork the repository
  2. Create a feature branch (git checkout -b feature/amazing-feature)
  3. Commit your changes (git commit -m 'Add some amazing feature')
  4. Add tests for new functionality
  5. Run the eval suite (python -m recall.evals.run_evals)
  6. Push to the branch (git push origin feature/amazing-feature)
  7. Open a Pull Request

Development Guidelines

  • Follow PEP 8 style guidelines
  • Write type hints for all functions
  • Add docstrings for public APIs
  • Update knowledge files when schema changes
  • Keep commits atomic and well-described
  • Test both backend and frontend changes

Developers

Dev GitHub
Keerthivasan S V Keerthivasan-Venkitajalam
Sri Krishna Vundavalli Sri-Krishna-V
Kavinesh Kavinesh11
Sai Nivedh SaiNivedh26

Report Bug · Request Feature

License

Apache License 2.0. See LICENSE for details.

Acknowledgments

Built with Agno MCP Framework. Inspired by OpenAI's in-house data agent. Learning Machine powered by pgvector. Built for the 2 Fast 2 MCP Hackathon.

About

Self-learning data agent delivering insights, not SQL. 6 layers of context grounding, autonomous learning, ChatGPT-style UI with real-time pipeline visualization. MCP-enabled. Built with FastAPI, PostgreSQL + pgvector, and Google Gemini. Production-ready with OpenTelemetry, Prometheus, and Docker.

recall-api.wonderfulsea-3100fdea.eastus.azurecontainerapps.io/

Topics

react docker kubernetes natural-language-processing postgresql prometheus gemini text-to-sql fastapi ai-agent pgvector data-agent

Resources

Readme

License

Apache-2.0 license
Activity

Stars

0 stars

Watchers

0 watching

Forks

0 forks
Report repository

Contributors

Languages

  • Python 70.4%
  • JavaScript 18.0%
  • Shell 5.4%
  • PowerShell 2.5%
  • PLpgSQL 2.5%
  • CSS 0.6%
  • Other 0.6%