one API to get all user desktop data
cross-platform, open source, 100% local,captures all desktop activities 24/7
how you can use it
publish your nextjs app
deploy your web app directly inside screenpipe desktop app
use as a backend library
integrate screenpipe as a backend for desktop context and activity
why use screenpipe?
perfect for building:
AI agents that need desktop context
apps requiring user activity tracking
applications using local LLMs
two main purposes:
backend for desktop context & user activity
nextjs apps directly to desktop - no wrappers
best use cases
ai agents
build AI assistants that understand user's desktop context and activities
search applications
create powerful search tools that index and query desktop activities
productivity tools
develop tools that analyze and optimize user workflows
what's included
desktop context & activity
records screen, audio, UI elements, and interactions 24/7
data processing
extracts text, transcribes audio, and stores in local sqlite db
ai capabilities
embeddings, RAG pipelines, local LLMs via Ollama, OpenAI/Anthropic proxy
app runtime
deploy your nextjs apps directly to desktop through screenpipe
what developers say
how apps are built
frontend
- • nextjs app with typescript
- • tailwind for styling
- • shadcn/ui components
- • deployed directly to desktop
backend
- • call screenpipe localhost:3030
- • store data on users device directly
- • call LLMs through local endpoints
- • create actions: clicks, typing
get started in minutes
1. install screenpipe library
add screenpipe to your Rust project
2. download screenpipe app
you'll receive the download link and occasional updates about screenpipe
get early access to the desktop app
3. create your app
bootstrap from our template
4. deploy it locally
run in screenpipe desktop
challenges we solved
native cross-platform complexities
managing different OS native APIs, permissions, and quirks across Windows, macOS, and Linux
performance challenges
optimizing CPU/GPU (when available) usage for continuous recording while balancing system resources and real-time processing
storage & processing
implementing video compression, structured data storage, and data deduplication
ocr & speech processing
integrating and optimizing OCR engines, handling multiple languages, real-time speech processing
vector search infrastructure
building efficient embedding storage/retrieval with HNSW indexes while optimizing for both speed and memory (WIP)
rust ecosystem complexity
managing FFI bindings, async runtime, unsafe code for system APIs, and cross-boundary error handling
get started
you'll receive the download link and occasional updates about screenpipe
for businesses
for investors