Building Blocks For a Better Web
IPFS is a constellation of open-source tools for content addressing. Build verifiable, portable, and resilient systems across public and private networks.
- 17.1K
- GitHub stars
- 1B+
- CIDs published
- 400k+
- public nodes
“It’s crucially important to have a distributed file system in our open hypertext system. IPFS is the missing piece that allows for a truly decentralized and open web.”
“Our local data infrastructure was stuck in customs. We set up IPFS on notebooks and a Raspberry Pi, and suddenly all scientists could sync, share, and collaborate, with automatic uploads back to Hamburg whenever the connection returned.”
“There is no central point of compromise anymore. There is no database that can be hacked, because the data just doesn’t exist, even in encrypted form, in one place anywhere. It only comes together at the edge device of the patient.”
“IPFS gave us a content-addressed foundation for publishing to the web, so our users can host and share sites that stay online without depending on any single server, and without the clunky UX that usually comes with decentralization.”
“IPFS is the standard for storing data in a verifiable and fully transparent way, thereby allowing us to build openly auditable governance systems.”
“It was such a relief to know I could store my videos in one place, each with its own IPFS URL and CID. Resilience is important to me, and having the work backed up means it’ll be around for a long time.”
Confidence in your data
Every tool in the IPFS constellation makes two commitments. Pick one to see it in motion, and grab the building blocks you need.
Hash around, and find out.
Running in science labs, game studios, art vaults, and low earth orbit.
Anywhere data needs to be verifiable, CIDs are helpful. A sample of what teams have built with the IPFS tools at the core.
Self-certifying data for the next wave of open networks.
AT Protocol uses CIDs so anyone on the network can verify what they receive. No trusted server in the middle.
Self-verifying records for open social
AT Protocol identifies every post, repo, and event by CID, so any client can verify what it received without trusting the server that delivered it.
Peer-to-peer hypermedia documents
Seed builds collaborative documents where every version, comment, and link is addressed by CID, durable across servers, editors, and time.
Offline-native knowledge graphs
Anytype uses content addressing to let users build personal knowledge webs that sync peer to peer, with no server lock-in.
Public, private, and everything in between
The same content-addressing model runs on the open internet or a closed network. Same CIDs, different routing.
Hybrid
Public peers, private data. Content is encrypted. CIDs route over the Amino DHT while ciphertext stays opaque to every transit node.
Private
All peers share a swarm key. Routing and transfer stay within the closed network. Nothing routes in or out without it.
Trustless Gateways
Gateways pull content from the IPFS network and serve it over HTTP. Browsers verify data against the CID locally, trust is rooted in CID, not the gateway.
Pick the tool that fits the job.
Start from the problem. Each row maps a real use case to the tools that solve it. Filter by language to narrow quickly.
Run Kubo with a swarm key for a closed IPFS network, or use iroh-blobs for lightweight QUIC-native transfer.
Run a Kubo or Helia node to pin and serve content; use ipfs-cluster for coordinated pinning across multiple nodes.
Upload to a pinning service, they handle replication, availability, and IPNI/DHT announcements.
Add ipfs-deploy-action to CI for a CID per build. Omnipin is a CLI deployment tool with ENS updates; SimplePage is a markdown-based publishing tool
Publish binaries, model weights, or data sets as signed CIDs bound to their source git commit, seedable over iroh-blobs, HTTP, IPFS, and BitTorrent.
Fetch and cryptographically verify IPFS content directly in the browser with no trusted gateway required.
Pack content-addressed blocks into a single CAR (Content Addressable aRchive) file for storage or transfer; verifiable and self-describing.
Serialize linked data with the DAG-CBOR and DAG-JSON codecs, the IPLD formats that give you deterministic encoding and CID links between blocks.
Represent files and directories as IPLD with UnixFS, the data model behind chunked files and folders on IPFS.
Check whether a CID is retrievable, inspect its DAG, and test gateways and IPNS records when content is not resolving as expected.
By the community, for the community.
IPFS is built in the open by 3,000+ contributors and counting. Join the disussions, track proposals, and find the next event online or near you.
