Threshold
Your context.
Your terms.
The infrastructure layer that lets people carry their own signals into every app they trust.
Feb 2026  ·  Integration Platform
The Problem
Every app asks for
permission like it
owns you.
Avg person: 6.7 platforms  ·  6.7 different identities  ·  6.7 permission dialogs
What's Actually Broken

You are fragmented
across your own tools

Your music

Trapped in Spotify

Your taste profile exists. Another app could use it to adapt to you — but asking you to re-enter it defeats the point.

Your work context

Trapped in project-control

Your focus state, your active project, your drift patterns. Computed locally. Invisible to every other app you use.

Your network

Trapped in LinkedIn

Your trusted collaborators' signals about shared work could help both of you — but there's no infrastructure for that exchange.

The Insight
What if you declared
what you trust —
instead of every app
claiming what it needs?
Same data. Different relationship.
Three Trust Primitives

Trust has three forms

You → App
"AI-DJ can read my Spotify and project focus data"
App gets a scoped token. You get an experience tuned to your actual context. Revoke in one place.
You → User
"Ryan can carry context about our shared work into apps we both use"
Trusted users bring your context to apps without you configuring anything. You get their network's signals back.
You → Network
"I'm part of this app's community — structural patterns can flow"
No individual data shared. Aggregate signals — overlapping taste, shared focus patterns — surface as recommendations.
The Key Differentiator

Apps don't just consume signals.
They produce them.

Source
Spotify
→
Raw data: top tracks, artists, listening history. App uses it directly via pass-through token.
Computes
AI-DJ
→
Derives: taste profile, focus anchors, drift state. These are new signals — not available anywhere else.
→
Publishes
focus_state
Subscribes
Any App
←
Another app you trust can subscribe to focus_state — without rebuilding drift detection. Your permission gates it.
From
Threshold
The UX

One familiar flow.
Different relationship.

App wants access
→
thresholdlabs.io/connect
→
You authorize
→
Token issued
It looks like OAuth. But the trust declaration is yours to initiate — not just to approve.

You can proactively say "I trust AI-DJ with my project focus data" before it even asks. When AI-DJ requests it, the auth is instant.

Revoke from one place. Propagates everywhere. No hunting through 8 different settings pages.
// What AI-DJ sees on redirect back
GET /callback?thld_token=thld_ct_...

// What the token grants
{
  "source": "project-control",
  "scopes": ["focus_state"],
  "user": "anon — never exposed",
  "revocable": true
}

// Not: your email, your ID, your data
// Just: the signal you said they can have
In Practice

Your music knows you're
working on the wrong thing.

You set "Events & Community" as your focus project.
You linked Bohemian Rhapsody as your anchor track.

You're actually deep in threshold code in iTerm.

project-control sees it. Pushes focus state to Threshold. AI-DJ subscribes — because you trusted it. Detects the mismatch.

Bohemian Rhapsody plays.

No companion script. No API key hunting. No configuration on the app side. Just: "I trust AI-DJ with my attention context."
What made this work

1 × Spotify connect token — AI-DJ reads your taste

1 × project-control grant — AI-DJ reads your focus state

1 × trust declaration — you said AI-DJ earns both

The app built no custom auth. The signals arrived. The experience happened.

The Platform Effect

What matures dissolves
into infrastructure.

Today

Custom auth,
every app

Every developer builds their own data access layer. OAuth flows, token refresh, scope negotiation. Solved problems, solved again.

Threshold

Signal layer
for everyone

Register your source. Declare your heuristics. Users trust you once. Every downstream app inherits the signal — with permission.

End state

Apps are mostly
configuration

Focus detection, taste profiling, context matching — platform primitives. New apps build experiences, not plumbing.

Landscape

No one else does
all three directions.

Platform User-centric trust Signal publishing User→User context Note
OAuth / OIDC NO NO NO App-centric. Reactive only. No signal layer.
Solid (Berners-Lee) YES NO PARTIAL Right model, no compelling first app. Cold start unsolved.
Farcaster / AT Protocol PARTIAL PARTIAL PARTIAL Protocol-first, content-only. Not a general signal layer.
IFTTT / Zapier NO NO NO Automation, not trust. App-centric wiring only.
Strava / Health Kit PARTIAL NO NO Vertical silo. Signals stay in ecosystem.
Threshold YES YES YES Trust-governed. Cross-app. Signal-native.
For Developers

Three ways to
build on Threshold

Consume a source

Pass-through
token

Register your app. Add a connect button. Users authorize from Threshold's UI. You get a scoped token. Use it directly against the source API.


Sources: spotify, github,
project-control, + more

Publish a signal

Heuristic
publisher

Declare what your app computes. Push updates via POST. Other apps can subscribe — with user permission. Your computation becomes a platform primitive.


focus_state, taste_profile,
sentiment, activity, + yours

Connect users

Trust graph
integration

Let users declare trust for each other inside your app. Their shared context flows bidirectionally. No individual data exposed — only structural signals.


User graph, shared context,
collective recommendations

Threshold
The platform that lets you be
the same person
in every app you trust.
THRESHOLD  ·  YOUR CONTEXT, YOUR TERMS
1 / 12
← → arrow keys