Connect on your terms.
Authorize a client, choose its tools and Channels, and share the exact context it may read. Access starts off and can be revoked.
Braid 2.0 preview · Your AI. Your agents. Your workspace.
Your local AI.
Your local AI.
Give it a home.
Start a conversation with your local model. Bring work from compatible AI clients into your Inbox through optional MCP connections. Review what becomes a note or handoff, connect approved rigs, and publish selected work with Folio.
Bundled Apple silicon Mac app · Windows/Linux Python launchers · Folio included
01Chat with your local AI.
Choose a model. No second rig needed.
02Review work from your agents.
Proposals arrive in your Inbox. You choose the next step.
03Publish with Folio.
Review and release selected work from the same app.
New in Braid 2.0 previewYour agents can bring
Your agents can bring
work home, too.
Optional MCP connections. An Inbox for proposals. Permissions and decisions stay with you.
Work arrives as a proposal.
Inspect its source and exact text. Edit an approved note locally, or stage a handoff. Sending is a separate human action.
See what connects.
Start with a clear List view. Open Network View to inspect rigs, models, MCP clients, Channels, and Folio outputs, using saved permissions and observed status.
One rig is enough. Local chat, readable streaming and code, Channels, approved rig sharing, and reviewed Folio publishing remain at the center. MCP client compatibility is still being qualified in this preview.
Your computers. Your models. Your project.Keep the useful context
Keep the useful context
when you change computers.
Decisions, constraints, research: share the parts your next conversation needs.
Carry the next useful step.
Review a project note and send it to an approved rig. See whether it was sent, received, accepted, or made available to chat. Free on your local network.
Ask with shared context.
Chat in Braid with relevant, permitted Channel notes. Inspect the exact material supplied to your local model.
Stay in charge.
Each rig chooses who can send, receive, and use its notes. Ordinary conversations and replies never auto-forward.
COMPUTER ONE
A project decision.
“Keep the second server within 500 credits.”
Review + shareApproved Braid Channel
COMPUTER TWO
A better next question.
“What’s our limit for the second server?”
Your local model · shared-note evidenceIllustrated workflow. Your selected local inference server generates the answer.
Folio publishing · Inside Braid
Private work.
Private work.
A public idea.
Turn a private conversation into a deliberate public edition. Folio keeps your original beside the public article. Choose an Editorial or Retro layout, add local images and rich formatting, and connect related editions.
- 01Try or choose. Start with Try a sample conversation — no model or second computer needed. Or use Publish with Folio on your own conversation or note.
- 02Shape and review. Choose passages, edit public excerpts, and add a byline, images, or related editions. Check the phone and print previews.
- 03Approve and export. Review local advisory highlights and approve the exact edition. Export HTML or a site ZIP; optional Classic HTML offers a simpler export using older web features.
Local advisory warnings highlight obvious file paths, private IPs, key-like strings, and machine hostnames for your review. Clear approval and published-state indicators show whether your draft matches the latest exported bundle or needs re-approval and re-export. Separately hosted pages are not checked. Nothing publishes without your explicit approval.
FOLIO / A SELECTED CONVERSATION
What should we
What should we
carry forward?
The useful part of a conversation can become a small, readable publication. Keep your voice. Choose what the next person sees.
FROM DRAFT TO EDITIONSelect · Edit · Review · Approve · Export
Local-first. Human-directed. Ready when you are.
Illustrative edition, not a public user transcript.
Start on one computerChoose a model.
Choose a model.
Start a conversation.
Run your model.
Start Ollama, vLLM, LM Studio, or a compatible server on this machine or your LAN.
Open a Channel.
In Braid, choose Chat with Local AI and open or create a Channel.
Choose your model.
Select your backend and completion model. Configure a server address if needed, then send your first message.
Grow from here.
Add approved rigs when you want to share context. Review selected conversations for publication with Folio.
One workspace. Your decisions.
Get Braid.
Get Braid.
Folio comes with it.
Your local AI workspace, with an Inbox for work from your agents and a publishing studio for the work you choose to share.
For local chat
Use your own Ollama, vLLM, LM Studio, or compatible local inference server and models. The Apple silicon Mac app includes Python and its app libraries. Windows/Linux launchers require CPython 3.10–3.13. Folio needs no model.
BRAID 2.0 PREVIEW · 2.0.0a2
The app, its runtime, and source
95.64 MB · Folio 0.4.0 included
Extract the ZIP. On an Apple silicon Mac, copy Braid 2.0 Preview.app to Applications and open it. On Windows, use Start Braid.cmd; on Linux, use sh "Start Braid.sh" with supported Python installed.
The preview opens a separate profile and can sit alongside 1.9.5. Choose Chat with Local AI for your model, or Connections to authorize an MCP client.
Download Braid 2.0 previewPreview qualification: the bundled Mac app is tested on macOS 15.5, Apple silicon. Windows/Linux launchers are included for target-host testing; native Windows/Linux runtimes are not included. Platform details.
Mac app: ad-hoc signed, not notarized. Model servers and model weights are configured separately. Prefer the existing 1.9.5 package?
Free local software. Optional membership.Start locally.
Start locally.
Add help and reach.
Folio is included in the free Braid app. Membership uses the same installation.
Braid + built-in Folio
Free
Approved LAN sharing, project Channels, local chat, and Folio publishing. Try a sample conversation before setting up a second rig.
Get Braid
FOUNDING MEMBERSHIPSupport + managed reach
$99 / year
Best-effort human help, early access, and provisioned Away from home for at least two device identities.
Join Founding MembershipManage billingHow membership and Away from home work
Managed WAN is early access and needs separate operator-assisted activation. Check availability with hello@intersignal.ai before joining for that capability alone. Checkout does not activate a rig.
Your home computer stays awake, online, and running Braid. Notes and remote Channel questions need separate peer permissions. Membership supplies neither computers nor hosted inference, models, or an unlimited service guarantee. Support is reasonable and best-effort, not a production SLA.
LAN and Folio stay free. You choose Managed WAN explicitly for approved destinations; a failed LAN handoff never silently switches to the internet. Detailed WAN setup and historical evidence.
The choices stay yours
A few useful answers.
Does Braid share all my conversations?
No. You select and review the material. Each receiver controls its own permissions. Private Folio originals and excluded passages stay out of Braid networking and public exports.
What does Folio publish?
The exact edition you review and explicitly approve. The split-view studio lets you include or exclude source blocks and redact public excerpts beside the article preview. Sensitive-detail warnings run locally and remain advisory. Editing public content invalidates approval. The state banner distinguishes local changes, an approved draft awaiting export, and content matching the latest immutable export; it does not verify your hosted website. Folio exports HTML or a site ZIP for your own hosting. Scheduled streams prepare review drafts. Choose Editorial or Retro layouts, add local images and related editions, and check phone or print previews. Classic HTML is a separate compatibility choice; live Braid connection invitations require Modern HTML.
Can a reader continue with me on Braid?
You may add an invitation to a publication. A reader requests a connection; you decide whether to accept. Rig-card exchange, fingerprint approval, and scope controls remain explicit. A public page never grants Braid authority.
Which model servers can I use?
Ollama models are discovered automatically. For vLLM, LM Studio, or another OpenAI-compatible local or LAN server, open Server settings, enter its address and port, and choose a model. No API key is needed unless that server requires one. Your existing inference server can keep running alongside its independent interface.
What can connected AI clients do?
With an explicit MCP grant, a client can discover its authorized rigs and Channels, read specifically selected context, propose a note or handoff, and inspect its own requests. Proposals wait for human review. Approval creates a local note draft or stages a handoff; sending remains a separate human decision. The MCP connection grants no remote execution or Folio publishing authority.
You can cancel an approved unsent handoff, revoke a client, or acknowledge an uncertain result without changing its delivery evidence. Completed work remains in Review history. The preview implements a tested MCP tools subset; named client products are not yet listed as verified integrations.
What happens to my existing Braid and Folio work?
Braid 2.0 Preview uses a separate profile and app identity. Keep 1.9.5 and its complete private profile as your working fallback. Existing identities, Channels, grants, and history are not automatically migrated into this preview. The ZIP includes setup, qualification, and profile guidance.
Verify the 2.0 preview download
Braid-2.0-Preview-2.0.0a2.zip · 95,641,886 bytes. SHA-256:
c90a1c36144f552a60db7df5d4757836aba20fde34b1ef30e75651d069f12530The ZIP contains the Apple silicon app and its runtime, companion CLI, public source, launchers, licenses, and qualification notes.
Which computers can use this preview?
The bundled native app is for Apple silicon Macs and has been exercised on macOS 15.5. Its configured minimum is macOS 13; other macOS versions and Intel Macs have not been qualified for this build. The app is ad-hoc signed, not notarized.
Windows and Linux use the included Python launchers, requiring CPython 3.10–3.13 with pip and venv. First setup needs internet or a prepared offline dependency pack. These are not bundled native Windows/Linux apps. The revised package still needs target-host and physical multi-rig/model qualification.
For the prior package, download Braid 1.9.5. Your local model server and model weights are separate from both packages.