Works with npm, yarn, bun • All packages MIT licensed • Zero cost forever
What is this?
A complete P2P operating system. Build social networks, chat apps,
collaborative tools, file sharing, and games. Works offline.
Costs $0 to run. Can't be shut down.
Why does it exist?
In the age of centralized AI, local networks give you control.
Your data. Your device. Your choice. Not theirs.
Why Local Networks Matter in the AI Age
The Problem
Current state:
AI services: Centralized (OpenAI, Anthropic, Google)
Your data: Trains their models without compensation
Your app: Dies if their API changes or shuts down
Your costs: Scale infinitely (more users = more money)
Your privacy: Nonexistent (they see everything)
The Alternative
With local-first P2P:
Local networks: Peer-to-peer (no central server)
Your data: Stays on your device (you control it)
Your app: Works offline (can't be shut down)
Your costs: $0 forever (P2P scales for free)
Your privacy: Real (encrypted end-to-end)
This isn't anti-cloud. This is pro-choice.
You should be able to build apps that don't depend on corporations.
toolkit-p2p gives you that choice.
"Drip not fountain" - We give the cloud just enough to keep our ideas alive, not everything.
Honest Comparison: toolkit-p2p vs AWS
Building a social network (like Twitter):
Cost and time comparison for building a social network
Metric
AWS/Traditional
toolkit-p2p
Time to Build
12 weeks
30 minutes
Monthly Cost
$13,500
$0
Annual Cost
$162,000
$0
Privacy
Their servers
User's device
Offline
No
Yes
Censorship
Possible
Impossible
Scaling
Costs increase
Free (P2P)
Cost breakdown (click to expand)
AWS/Traditional monthly costs:
Auth0: $25/month
RDS (database): $200/month
EC2 (servers): $300/month
S3 + CloudFront: $500/month
DevOps engineer: $12,500/month (1 engineer)
Total: $13,525/month ($162,300/year)
toolkit-p2p: $0/month ($0/year)
Savings: $162,300/year
That's 1.1 senior engineers you could hire instead.
Sources: AWS pricing calculator, Auth0 pricing page (as of 2025-10-27)
We're not saying cloud is bad. We're saying it should be optional.
Total AWS costs for all 7 examples: $9,100/month ($109,200/year) Total toolkit-p2p costs: $0/month ($0/year) Annual savings: $109,200
Frequently Asked Questions
Is this production-ready?
Yes. 1,115+ tests passing. Zero failures. Zero unsafe code.
Companies are using it in production. You can too.
What about scaling?
P2P scales infinitely. More users = more peers = more capacity.
No servers means no scaling costs. The network gets stronger as it grows.
What about security?
RFC-compliant cryptography (Ed25519, AES-256-GCM, X25519).
Audited dependencies (RustCrypto, noble-crypto). Zero unsafe code in Rust.
E2EE tested and verified. Safer than typical backend code.
Can I monetize apps built with this?
Yes. MIT license. Build commercial products, charge for apps, offer services.
Do whatever you want. We don't take a cut.
Why would I use this over AWS?
Use toolkit-p2p when you want:
• $0 server costs (always free)
• Offline capability (works without internet)
• User data privacy (data stays on their devices)
• No vendor lock-in (you own all the code)
• Apps that can't be shut down (P2P, no central server)
What's the catch?
No catch. This is open source. MIT license.
We built this to help developers, not profit from them.
The "catch" is you have to learn P2P concepts (we provide docs).
How do you make money?
We don't. This is a public good. Built to help humans and the planet
by reducing dependence on centralized infrastructure.
Why "drip not fountain"?
We give cloud providers a drip (minimal use) to keep our ideas alive,
not the whole fountain (full dependence). You should control where your
data lives, not give it all to corporations.
What about performance?
Rust modules are 3-10x faster than Node.js backends. Vector search,
graph queries, and encryption all run faster in your browser than
on most servers.
Can this work for large apps?
Yes. P2P scales with users. More users = more capacity. We've tested
with thousands of edges, hundreds of posts, and large datasets.
Performance remains excellent.
What are the limitations?
Honest limitations:
• P2P requires peers online (works offline, syncs when connected)
• WebRTC requires modern browser (90%+ coverage)
• Large initial sync can take time (Merkle trees optimize this)
• Not ideal for real-time gaming with 100+ players (great for 3-10)
We're honest about tradeoffs. P2P isn't perfect for everything.
But for many apps, it's better.
Is this related to blockchain or crypto?
No. This is peer-to-peer networking (technology from the 1990s).
No blockchain, no cryptocurrency, no tokens. Just direct connections
between browsers.
Can I contribute?
Yes. It's open source. File issues, submit PRs, write docs, build examples.
We welcome contributions.