Harper

# Harper Harper is a single in-process runtime that unifies your database, application logic, APIs, caching, real-time messaging, and agent loops, replicated across a multi-region data application fabric. Because data sits in the same process as your code, Harper delivers roughly 1ms data retrieval, up to 85% fewer LLM and origin calls, and 30 to 70% faster page loads. Harper is built for developers and AI coding agents. You define a schema, and Harper automatically exposes REST and GraphQL APIs, vector search, blob storage, caching, and pub/sub over Server-Sent Events, WebSocket, and MQTT, all from one runtime with one deploy target. The same code runs on a laptop, in managed cloud, at the edge, and on-prem. ## Getting started - [Full Technical Documentation](https://docs.harperdb.io/llms.txt): Agent entry point for learning about the technical details of building with Harper. - [Install Harper](https://docs.harperdb.io/learn/getting-started/install-and-connect-harper): Install the runtime locally via npm, Docker, or a package manager. - [Quick Start](https://docs.harperdb.io/learn/getting-started/create-your-first-application): Stand up your first Harper application and hit a live API in minutes. - [create-harper](https://github.com/HarperFast/create-harper): One command scaffolds a Harper project with schema, resources, config, and agent onboarding skills files. ## Core capabilities - [Applications](https://docs.harperdb.io/learn/developers/harper-applications-in-depth): Build full-stack backends with JavaScript classes that run in-process alongside your data. - [Defining Schemas](https://docs.harperdb.io/reference/v5/database/schema): Define tables in GraphQL to auto-generate REST and GraphQL APIs, with typed and indexed fields. - [REST](https://docs.harperdb.io/reference/v5/rest/overview): Query, filter, sort, join, and paginate directly from a URL or a fetch call. - [Caching](https://docs.harperdb.io/learn/developers/caching-with-harper): Cache external sources with configurable TTL and stampede protection, with no separate cache tier. - [Real-Time](https://docs.harperdb.io/reference/v5/rest/websockets): Subscribe to record changes over Server-Sent Events, WebSocket, or MQTT, with no separate broker. - [Vector Indexing](https://docs.harperdb.io/reference/v5/database/schema#vector-indexing): Store embeddings beside your data with an HNSW index and run semantic search through the same API. ## Deploy - [Harper Fabric](https://docs.harperdb.io/fabric): Deploy your entire backend to a globally distributed network with a single command. - [Sign up for Fabric](https://fabric.harper.fast/#/sign-up): Create a Fabric account and start building, with no credit card required. - [npm package](https://www.npmjs.com/package/harper): Install the Harper runtime from npm. - [Docker images](https://hub.docker.com/u/harperfast): Run Harper in a container. ## Open source and examples - [Harper source code](https://github.com/HarperFast/harper): The open source Harper unified runtime. - [All apps and plugins](https://github.com/search?q=topic%3Aharper-app+org%3AHarperFast+&type=repositories&s=stars&o=desc): Templates and plugins that accelerate new builds. - [Cortex](https://www.harper.fast/ai/cortex): Open source shared memory layer for AI tools and agents, with far higher recall capacity than alternatives. - [MCP Server](https://github.com/HarperFast/mcp-server): Expose Harper data as Model Context Protocol resources for AI tools. - [Semantic Cache Example](https://github.com/HarperFast/semantic-cache-example): A vector-based cache that cuts redundant LLM calls. - [Next.js plugin](https://github.com/HarperFast/nextjs): Run and develop Next.js apps on Harper. ## Product and use cases - [Run Agents](https://www.harper.fast/ai/runtime): A runtime for agents with data, semantic cache, real-time, and REST interfaces in one process. - [B […gekürzt]