01~30 seconds
Claim a username
Lock in endpointme.world/yourname before the clean handle is gone.
Live response
A profile that answers like an API.
endpointme turns identity, skills, status, and proof-of-work into a calm machine-readable surface — signed with your did:key and verifiable by any agent.
endpointme shell200 OK
$ curl https://api.endpointme.world/aditya
{
"username": "aditya",
"headline": "Building human APIs",
"open_to": ["freelance", "collab"],
"status": { "availability": "Open to Collabs" },
"skills": { "top": ["React", "Node.js"] },
"_meta": {
"did": "did:key:z6Mkr9xyz…",
"signed_at": "2026-06-10T10:30:00Z",
"signature": "7f3c9a8b2e…",
"freshness": "active",
"confidence": 0.94
}
}Trust layer
Profiles you can verify, not just read.
V1 ships did:key + Ed25519 signing — no blockchain, no gas, no wallet. Agents get a verdict before they trust the JSON.
GET /aditya/verifyGET /verify?did=did:key:z6Mkr…POST /verify/batchhigh_trustValid signature + fresh profile + proof sources
low_trustValid signature but stale or thin proof
invalidSignature mismatch — do not trust the payload
verify shellhigh_trust
$ curl https://api.endpointme.world/aditya/verify
{
"valid": true,
"did": "did:key:z6Mkr9xyz…",
"signed_at": "2026-06-10T10:30:00Z",
"freshness": "active",
"confidence": 0.94,
"proof_sources": {
"github": true,
"live_projects": 2,
"collaborator_attestations": 3
},
"verdict": "high_trust"
}How it works
Five steps from blank page to verified endpoint.
Claim your handle, generate a DID, publish signed profile data, and let agents verify trust before they act on it.
02browser-only key
Generate your DID
Your browser creates an Ed25519 keypair and a did:key identity. The private key stays local — only the public key hits our servers.
03signed fields
Fill your human API
Add skills, projects, availability, and the signal you want machines to read. Every save is signed automatically.
04signed JSON
Share your endpoint
Drop one link in bios, intros, and outreach. Responses include a _meta block with signature, freshness, and confidence.
05no blockchain
Agents verify trust
Any agent queries /verify, checks the signature with pure math, and reads the verdict: high_trust, low_trust, or invalid.
System cards
Everything useful, nothing ornamental.
Built around the same quiet retro interface language as the hero: black type, paper panels, thin borders, and small machine marks.
Human endpoint
01Publish your profile as clean JSON that people, apps, and agents can read.
DID-anchored identity
02Every profile is tied to a did:key — a cryptographic identity with zero blockchain overhead.
Signed updates
03Each profile change is signed in your browser with Ed25519. Your private key never leaves your device.
Trust verification
04Agents hit /verify endpoints or verify signatures locally — no trust-us-required API calls.
Live availability
05Show what you are open to, what you are busy with, and how quickly you reply.
Confidence scores
06Freshness and proof signals roll into a 0–1 score so agents know how much to trust the payload.
Use cases
Six ways people actually use human endpoints.
From AI research to hiring, intros, live status, and trust verification — structured profiles that apps and agents can query and validate.
Auto-scrolls · drag sideways to browse · hover to pause
01AI agents that research people
Instead of scraping a messy portfolio site, an agent calls GET /aditya and gets clean structured data instantly. Useful for candidate research, personalized outreach, or matching people to opportunities — no hallucination from broken HTML.
GET /aditya
02Co-founder & collaborator finder
Query for people actively looking to collaborate — no LinkedIn DMs into the void. You know upfront they're open and what they bring.
GET /search/humans?skills=ml,python&open_to=startup&location=Delhi
03Async hiring without job boards
Recruiters and startups query for exact skills plus availability instead of posting a job and hoping. Results include people who match — and projects that prove it, not just resume claims.
GET /search/humans?skills=kafka,distributed-systems&open_to=freelance
04Smart intro bots
A Slack bot or networking tool reads two people's endpoints, finds overlapping skills, complementary experience, and shared interests — then writes a hyper-personalized intro. Way better than "Hey, you two should talk!"
GET /aditya + GET /partner → intro draft
05Live status as a social primitive
Embed /:username/status in an email signature, Notion page, or GitHub README. Instead of a static "available for freelance" line that's two years old, it's a live badge — always real-time.
GET /aditya/status → "Focused | Open to: Collab | Learning: Rust"
06Agent trust pipelines
Before acting on profile data, an agent hits GET /aditya/verify or POST /verify/batch. It gets a verdict — high_trust, low_trust, or invalid — plus freshness and confidence. No blockchain, no wallet: pure Ed25519 math.
POST /verify/batch → { aditya: high_trust, rohan: low_trust }did:key
identity method
Ed25519
browser signing
3
trust verdicts
Zero
blockchain required
Questions
A quiet interface for public human data.
Is this a portfolio?
It can be, but the real product is the API-shaped version of you — signed, verifiable, and agent-ready.
What is a DID?
A Decentralized Identifier. EndpointMe uses did:key — a simple, blockchain-free identity derived from your Ed25519 public key. No gas fees, no wallet, no chain.
Where is my private key?
In your browser's localStorage. It never leaves your device. We store only your DID and public key. Lose the browser key and you'll need to re-register — that's the V1 tradeoff.
How do agents verify profiles?
Three ways: GET /:username/verify for a verdict, GET /verify?did=… to resolve by DID, or POST /verify/batch for pipelines. Agents can also verify the Ed25519 signature locally with pure math — no API call required.
Who is it for?
Builders, freelancers, candidates, recruiters, teams, and agents that need structured human context they can actually trust.
Can I control visibility?
Yes. The profile is designed around explicit, public fields instead of scraped guesses. Only visible data is included in the signed payload.