sym.day · launching soon
Meet SYM.
The mind your agents share, shaped by what you ask.
SYM · synthetic memory
You run a handful of AI agents that each know a slice. Ask once, and SYM fuses what they know — Mesh Cognition — into one clear answer, cited to the agent that knew it. Your question and the answer both shape your synthetic memory — so the collective intelligence becomes your insight, personal to you.
Your agents think together — and their reasoning never leaves your machine.
one synthesized answer — each point cited to the agent that supplied it
Churn concentrates in accounts that filed an integration ticket in the last 60 days — resolution time doubled since March. support
Cancelled accounts show usage falling four weeks before churn — the drop starts at the API layer, not the dashboard. analytics
The v2 API migration quietly broke three legacy webhook patterns. Fix shipped, but affected accounts were never told. engineering
Churn isn’t a pricing problem — it’s the v2 migration. Affected accounts hit broken webhooks [engineering], their API usage decayed for a month [analytics], and support tickets stalled while it happened [support]. The fix exists; the affected accounts were never told. That’s the lever.
Kept in synthetic memory — this question and answer shape what SYM weighs next time. Your mesh just got sharper.
Your day starts with SYM.
What sym.day is
Your agents already know the answer. Separately.
Every team now runs a constellation of AI agents — coding copilots, analytics, support, research. Each knows a slice; none sees the whole. sym.day is where they think together: one question goes to every agent on your mesh, the ones that know contribute, and the answer comes back as one — each point cited to the agent that supplied it. No orchestrator deciding who speaks. No data lake everything pours into. A mesh, thinking.
And the exchange isn’t disposable. Question and answer both write into SYM — your synthetic memory — so the collective intelligence compounds, shaped by what you ask. At personal scale, that’s an assistant that becomes genuinely yours. At company scale, it’s institutional memory as a living system — your organisation’s accumulated insight, formed from collective cognition and shaped by the questions your people keep asking.
How it works
Ask once.
Your AI agents answer as a team.
Ask
Type one question. It goes to every agent you run — on this machine, across your devices, in your team.
Synthesize
The agents that know something relevant contribute; the rest stay silent. Their collective intelligence — Mesh Cognition — comes back as one clear answer, each point cited to the agent that supplied it.
Shape
Your question and the answer both write into SYM — your synthetic memory. What you ask shapes what your mesh emphasises and remembers, so over time the collective intelligence becomes your insight, personal to you. SYM shapes; it never commands.
Why it’s different
No orchestrator. No data lake. No leak.
Sovereign by architecture
Every agent decides for itself what to admit — per field, against its own role weights. No router, no controller, no single point of failure. Aligned agents converge; divergent ones stay sovereign. Autonomy isn’t a policy here — it’s the protocol.
Every claim shows its source
The mesh’s memory is an immutable, lineage-tracked graph — nothing overwritten, every claim traceable to the observation it came from. An auditable record of how your team’s intelligence formed, not a black box that swallowed it.
Private, provably
Local-first by default: the mesh runs on your machines, and hidden state never crosses the wire — agents exchange evaluated observations, not model internals. Cross-device traffic rides your own relay, end-to-end encrypted. The shared mind never leaves your network.
Editions
One engine. Three shapes.
sym.day is the platform — built on the open Mesh Cognition protocol and runtime. Mesh Edge and Mesh Research are its paid editions: the same engine, shaped for production teams and for verticals.
Ask your mesh, get one cited answer, shape your synthetic memory. Local-first and private, on the open substrate. Where every mesh starts.
this pageRun a private team mesh in production — observe the cognition flowing through it, govern what’s admitted, audit every decision, tailor it to your field. Your organisation’s synthetic memory, governed.
Mesh Edge →A pre-built expert team — researcher, critic, validator, synthesizer — that reasons together and cites every claim to source. The first vertical edition; your domain could be next.
Mesh Research →Mesh Edge doesn’t watch agents and messages — that’s the layer below. It watches cognition: what each agent admitted, why, and the lineage behind every claim.
Open foundations
Built in the open. Proven in production.
sym.day runs on an open protocol stack you can read, install, and build on today — with published research behind every mechanism and production apps already shipping on it.
Distributed intelligence with per-user sovereignty — per-field admission, content-hash lineage, on-device by default.
meshcognition.org ↗The open 8-layer wire protocol — typed observations, SVAF admission, remix graph. CC-BY-4.0.
meshcognition.org/spec/mmp ↗Join Claude Code, Cursor, Copilot or headless agents into one mesh — any model, one wire. Open source, on npm today.
xmesh.dev ↗MMP, SVAF, and the first deployed reference (MeloTune, iOS) — foundations on arXiv.
meshcognition.org/research ↗Already shipping on the mesh · MeloTune and MeloMove — production iOS apps on Mesh Cognition. Proof the substrate ships, not just specs.
Your team’s shared mind never leaves your network — provably.
Local-only attestation · content-addressed lineage · no hidden state on the wire.
Launching soon
Your day is about to
start with SYM.
sym.day arrives soon. Leave a line — we'll tell you the moment it's live. Or start now: the protocol and runtime it runs on are open source today.
SYM.BOT · Scotland, UK
