All features

Alyph · Threads

Every workstream in its own thread. Switching costs nothing.

Solo founders and small teams don't have one conversation — they have ten. Client work, admin, coding, life admin: Alyph keeps each in its own context, so switching between them costs nothing.

The short answer

Threads give every project, client, and workstream its own isolated context on the canvas. Contexts never mix — your pricing thread never inherits your tax question — so switching between workstreams costs nothing.

09:00 · client work

Smith & Co: the pricing page.

Claude 3.5 Sonnet

Annual toggle first. Two variants.

12:30 · admin

Invoice #18 is 40 days out.

ChatGPT 5.6 Sol

Follow-up draft. Firm, warm, one page.

15:00 · coding

Vite segfaults on Macbook M3.

Claude 3.5 Sonnet

Known issue. Pin esbuild to 0.21.

21:00 · life

Apollo, vet claim #4412

2 documents

hospital-letter.pdf

What you get

  • One thread per client, project, or workstream
  • Files stay with the work they belong to
  • Work and life admin never bleed into each other
  • Pick up exactly where you left off
  • Branch inside any thread when it gets hairy

The many-hats problem

Client work at nine. Admin at noon. Coding at three. The dog’s insurance claim at nine. In a linear chat app, that’s either ten half-broken conversations you can’t find, or one long one where the pricing logic keeps accidentally importing tax advice.

How threads stay clean

Each thread holds its own context: its files, its decisions, its history. The model only ever reads the thread you’re in (plus the shared context you chose), so there’s nothing to untangle when you switch. And when a thread sprouts a tangent, that’s what branches are for.

Honest limits

Canvases currently get sluggish around 100+ active cards, and rendering optimizations are ongoing. In practice: prune dead branches and archive finished projects, and the canvas stays fast. We’d rather tell you the number than have you find it.

Questions, answered

How many threads can a canvas hold?

There’s no hard cap on threads; the practical limit today is visual: canvases get sluggish around 100+ active cards. Pruning finished branches keeps a busy canvas fast, and we’re actively optimizing rendering.

Can threads share files?

Yes. Files placed at the root are shared context: every thread and branch below inherits them. Project-specific files stay inside their own thread. How files work →

Is this just folders with a nicer name?

No. Folders organize where things live; threads control what the model sees. Isolation is the point: context from one workstream never shapes answers in another.

Try threads on the canvas

Free to start, with $5 of welcome credit. No subscription — cancel nothing, ever.