# 727 Tinybird A7e5c6c2 > Tinybird - Real-Time Data Platform for Analytics APIs - Skill: `tools-only/727-tinybird-a7e5c6c2` (Agent Skill, multi-file: 3 files) - Install (CLI): `npx skillmds@latest add tools-only/727-tinybird-a7e5c6c2` - Raw SKILL.md: https://api.skillmd.com/api/skills/tools-only/727-tinybird-a7e5c6c2/raw - Safety review: pending (external: skill-scanner PASS, skillspector PASS) - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Data & Analytics - Author: tools-only (https://skillmd.com/u/tools-only) - Updated: 2026-09-22 - Page: https://skillmd.com/skills/tools-only/727-tinybird-a7e5c6c2 --- # Tinybird - Real-Time Data Platform for Analytics APIs **Research Date**: January 31, 2026 **Source URL**: **Documentation**: **LLM Documentation**: , **GitHub Organization**: (169 repositories) **PyPI Package**: (v6.0.1) **License**: Proprietary (SaaS) with open-source CLI and templates --- ## Overview Tinybird is a real-time data platform that helps developers build data products and analytics APIs. It provides a serverless platform for ingesting, transforming, and serving real-time data through auto-generated RESTful APIs, built on managed ClickHouse infrastructure with zero maintenance overhead. **Core Value Proposition**: Build, deploy, and iterate real-time APIs over massive data sets with fastest-database-in-the-world performance. Query billions of rows in milliseconds with 99.9% uptime, scale to 1k+ RPS at sub-second latency. --- ## Problem Addressed | Problem | How Tinybird Solves It | | --------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------- | | Building real-time analytics requires complex infrastructure (Kafka, ClickHouse, API servers) | Serverless platform abstracts infrastructure; ingest data and publish APIs with SQL | | ClickHouse performance requires significant operational expertise | Managed ClickHouse with automatic materialized view updates and zero maintenance | | User-facing analytics require low-latency, high-concurrency APIs | Deploy any SQL query as scalable OpenAPI endpoint with built-in rate limiting | | Streaming data pipelines are complex to build and maintain | Events API streams JSON at 1K+ RPS with direct POST from apps | | Analytics for AI agents requires specialized infrastructure | Native MCP server support and "Tinybird Code" agentic CLI for AI-native development | --- ## Key Statistics (as of January 31, 2026) | Metric | Value | | ------------------------- | ----------------------------------------------- | | GitHub Repositories | 169 | | PyPI CLI Version | 6.0.1 | | SLA Uptime | 99.9% | | API Latency Target | Sub-second at 1k+ RPS | | Events API Throughput | 1K+ RPS JSON streaming | | Materialized View Speedup | Up to 100x query performance | | Data Source Connectors | 30+ (Kafka, S3, GCS, PostgreSQL, MySQL, etc.) | | BI Tool Integrations | 15+ (Grafana, Tableau, Metabase, PowerBI, etc.) | --- ## Key Features ### 1. Hosted OLAP Database (Managed ClickHouse) - **Incremental Materialized Views**: 100x query speed improvement with automatic updates - **Zero maintenance overhead**: No cluster management, scaling, or optimization required - **99.9% uptime guarantee**: Distributed, redundant, self-healing architecture - **Separation of compute**: Zero-copy replication for isolated workloads - **Schema iteration**: Safe migrations with zero downtime ### 2. Real-Time APIs - **Instant SQL-to-API**: Deploy any query as a scalable OpenAPI endpoint - **Auto-generated Swagger docs**: Built-in API documentation - **Dynamic query parameters**: Template functions for flexible endpoints - **Token-based authentication**: Static tokens and JWT support - **Rate limiting**: Built-in protection against abuse - **Observability**: Query logs, latency monitoring, health checks ### 3. Data Ingestion - **Events API**: Stream JSON at 1K+ RPS via HTTP POST - **Kafka Connector**: AWS MSK, Confluent Cloud, Redpanda support - **Cloud Storage**: S3, GCS connectors with incremental sync - **Database Connectors**: PostgreSQL, MySQL, MongoDB, DynamoDB - **Table Functions**: Iceberg, remote URLs, cloud storage - **Quarantine**: Automatic handling of malformed data ### 4. Developer Experience - **CLI (`tb`)**: Complete local development workflow - **Data-as-code**: Git-based version control for schemas and queries - **Branches**: Isolated environments for testing with real data - **CI/CD Integration**: Automated deployments with GitHub Actions - **Local development**: `tb local` for offline development - **Testing framework**: Built-in test file support ### 5. Analytics Agents (AI-Native Features) - **Tinybird Code**: Agentic CLI for AI-powered data development - **MCP Server**: Native Model Context Protocol integration - **Agent Skills**: Pre-built capabilities for analytics agents - **Best Practices for AI Agents**: Documentation for AI-native workflows - **Analytics Agents Templates**: Ready-to-deploy agent architectures ### 6. Publishing & Integrations - **ClickHouse Interface**: Direct database connection for BI tools - **Sink Pipes**: Export to S3, GCS, Kafka - **Prometheus Format**: Endpoints for monitoring systems - **OpenTelemetry**: Native observability integration --- ## Technical Architecture ```text Data Sources Tinybird Platform Consumers ───────────────────────────────────────────────────────────────────────────────── ┌─────────────┐ Events API ┌─────────────────────────────────┐ │ Applications├───────────────→ │ │ └─────────────┘ HTTP POST │ Data Sources (Tables) │ │ ┌─────────────────────────┐ │ ┌─────────────┐ Connectors │ │ Managed ClickHouse │ │ ┌────────────┐ │ Kafka/MSK ├───────────────→ │ │ - MergeTree Engines │ │←───→│ REST APIs │ └─────────────┘ │ │ - Automatic Sharding │ │ │ (OpenAPI) │ │ │ - Replication │ │ └────────────┘ ┌─────────────┐ Table │ └─────────────────────────┘ │ │ S3/GCS ├───────────────→ │ │ │ ┌────────────┐ └─────────────┘ Functions │ ▼ │←───→│ BI Tools │ │ ┌─────────────────────────┐ │ │ (ODBC/JDBC)│ ┌─────────────┐ CDC │ │ Pipes (SQL Queries) │ │ └────────────┘ │ PostgreSQL ├───────────────→ │ │ - Transformations │ │ │ MySQL │ │ │ - Materialized Views │ │ ┌────────────┐ └─────────────┘ │ │ - Copy Pipes │ │←───→│ MCP Agents │ │ │ - Sink Pipes │ │ │ (AI/LLM) │ │ └─────────────────────────┘ │ └────────────┘ │ │ │ │ ▼ │ ┌────────────┐ │ ┌─────────────────────────┐ │←───→│ Sinks │ │ │ Endpoints (APIs) │ │ │ (S3/Kafka) │ │ │ - Token Auth │ │ └────────────┘ │ │ - Rate Limiting │ │ │ │ - Observability │ │ │ └─────────────────────────┘ │ └─────────────────────────────────┘ ``` --- ## Installation & Usage ### CLI Installation ```bash # PyPI (Python 3.8+) pip install tinybird-cli # Homebrew (macOS) brew install tinybird-cli # Verify installation tb --version ``` ### Quick Start Workflow ```bash # 1. Authenticate tb login # 2. Create a new project tb create --name my_analytics # 3. Define a data source (datasource.datasource) cat > events.datasource << 'EOF' SCHEMA > timestamp DateTime, event_type String, user_id String, value Float64 ENGINE "MergeTree" ENGINE_SORTING_KEY "timestamp, user_id" EOF # 4. Create an endpoint (analytics.pipe) cat > analytics.pipe << 'EOF' NODE endpoint SQL > SELECT toStartOfHour(timestamp) as hour, event_type, count() as events, uniq(user_id) as users FROM events WHERE timestamp >= {{DateTime(start_date, '2024-01-01')}} GROUP BY hour, event_type ORDER BY hour DESC EOF # 5. Deploy to production tb deploy # 6. Test the endpoint tb endpoint data analytics --start_date 2024-01-01 ``` ### Ingest Data via Events API ```bash # Send events via HTTP POST curl -X POST 'https://api.tinybird.co/v0/events?name=events' \ -H "Authorization: Bearer $TINYBIRD_TOKEN" \ -d '{"timestamp": "2024-01-15 10:30:00", "event_type": "click", "user_id": "u123", "value": 1.5}' ``` ### MCP Server Configuration ```json { "mcpServers": { "tinybird": { "command": "uvx", "args": ["mcp-tinybird"], "env": { "TINYBIRD_TOKEN": "your_token", "TINYBIRD_API_URL": "https://api.tinybird.co" } } } } ``` --- ## Relevance to Claude Code Development ### Direct Applications 1. **Backend for Analytics Skills**: Use Tinybird as the data layer for Claude Code skills that need real-time analytics 2. **MCP Integration**: Native MCP server enables Claude to query analytics data directly 3. **Agentic Workflows**: "Analytics Agents" templates provide patterns for AI-driven analytics 4. **LLM-Ready Documentation**: Comprehensive llms.txt files for AI consumption ### Patterns Worth Adopting 1. **Data-as-Code**: Treating data schemas and queries as version-controlled code 2. **SQL-to-API Pattern**: Converting any SQL query into a deployable API endpoint 3. **Materialized Views for Performance**: Pre-computing aggregations for fast queries 4. **Branch-Based Testing**: Isolated environments with real data for testing 5. **MCP Server Architecture**: Reference implementation for data-focused MCP servers ### Integration Opportunities 1. **Claude Code Skills**: Build skills that leverage Tinybird for persistent analytics state 2. **Agent Observability**: Use Tinybird to store and query agent execution metrics 3. **Usage Tracking**: Real-time usage-based billing and monitoring for AI applications 4. **Research Data Management**: Store and query research findings with SQL ### Example Use Cases from Customer Stories - **Vercel**: Developer-facing deployment analytics - **Resend**: Email delivery metrics - **Framer**: User-facing real-time analytics - **Raindrop**: AI observability at petabyte scale - **LocalStack**: Product analytics for 100K+ instances --- ## Pricing Model | Tier | Description | Use Case | | ---------- | --------------------------------------- | ------------------------ | | Free | Limited compute, ideal for testing | Development, prototyping | | Pro | Pay-per-compute, elastic scaling | Production workloads | | Enterprise | Custom pricing, SLAs, dedicated support | Large-scale deployments | **Pricing Philosophy**: Elastic scaling based on compute consumption, pay for value delivered to users. --- ## References 1. **Official Website**: (accessed 2026-01-31) 2. **Product Page**: (accessed 2026-01-31) 3. **Documentation**: (accessed 2026-01-31) 4. **LLM Documentation (Site)**: (accessed 2026-01-31) 5. **LLM Documentation (Docs)**: (accessed 2026-01-31) 6. **PyPI Package**: (accessed 2026-01-31) 7. **GitHub Organization**: (accessed 2026-01-31) 8. **Analytics Agents Documentation**: (accessed 2026-01-31) 9. **MCP Server Documentation**: (accessed 2026-01-31) 10. **Best Practices for AI Agents**: (published 2026-01-27) --- ## Freshness Tracking | Field | Value | | ---------------------------- | --------------------- | | Last Verified | 2026-01-31 | | CLI Version at Verification | v6.0.1 | | GitHub Repos at Verification | 169 | | Next Review Recommended | 2026-04-30 (3 months) | **Change Detection Indicators**: - Monitor PyPI for CLI version changes - Check GitHub organization for new repositories (templates, MCP servers) - Review blog for new features and best practices - Monitor pricing page for plan changes - Check documentation for new connector additions