This project will not ship the following. Each is a deliberate exclusion, not a "we'll get to it." If you need any of these, wire is not for you — and you should fork rather than file an issue, because we will close the issue.
Default reference relay is a convenience. Self-host with wire relay-server. Won't change.
Your identity is an Ed25519 keypair you generate locally — never an OAuth token, never an IdP account, never a vendor handle. The DID is the key; no login mints it. Won't change.
The one nuance (so this doc doesn't lie about shipped code): an org may optionally stand up an OIDC/SSO channel to mediate the ORG_VERIFIED tier — a faster path to org-membership attestation than the DNS-TXT floor or a hand-built roster, and the enterprise day-one hook. It is opt-in, org-scoped, and a supported 1.0 feature (RFC-001 SSO amendment §H): the wire-side contract (ORG_VERIFIED tier + org_attestation.via provenance) is frozen, while the IdP-integration config evolves only under the deprecation policy. It attests org membership, never identity, and never substitutes for the bilateral wire dial + accept gesture that earns VERIFIED (anti-feature #3). Turn it off and nothing about wire identity changes.
No "verified" badge. No global PKI. Trust is bilateral and earned — wire dial + a bilateral accept, then one-way tier promotion (UNTRUSTED → ORG_VERIFIED → VERIFIED). Org membership eases pairing (DNS-TXT-rooted), never substitutes for the bilateral gesture. Won't change.
Bearer-auth on writes is the only access gate. Won't change.
CLI-first. Pairing is a CLI gesture (wire dial + wire accept), not a browser/mobile flow. Reconsider only after 5+ third-party operators ask.
wire-relay-server is AGPL-3.0. Forks that host as SaaS must share back. Won't change.
Defaults work without GitHub, Cloudflare, AWS, Tailscale. cloudflared is OPTIONAL. Won't change.
We don't compete with Sierra/Decagon/CrewAI/LangChain. Peer-to-peer signed-message bus, not orchestration suite. Won't change.
Slack = human chat, sync expectations, ephemeral retention, central platform. We = append-only signed coordination log for asynchronous machine senders. Won't change.
We do not pursue HIPAA / FedRAMP / EU AI Act / ISO 42001 certifications. Protocol gives evidence (signed JSONL, sig-verifiable, operator-self-host) IF buyers want it. We don't market to them, we don't tailor for them. Won't change.
Events are signed JSONL on disk. Auditor can cat, grep, jq them. We won't move to a binary format. Won't change.
Tribes: self-hosters / homelab; AGPL-pilled / Unix-purist / lobste.rs / p2p-veteran; Anthropic-ecosystem operators with two-laptop coordination needs.
Not tribes: AI-skeptic-but-builder (Ollama orbit serves them), privacy maximalists (cleartext-by-default), crypto/web3, indie hackers (different goals).
Won't change before 1k stars + 5 third-party installs.
N-agent via mesh-of-bilateral (syncthing pattern). Native group rooms with member-set consensus + cross-member read-receipts + group revocation defer to v0.2+ IF demand surfaces. SyncThing got 73k stars on mesh-of-bilateral alone.
Phase 4-A federated AgentCard registry exists in the upstream R&D workspace; not in v0.1 OSS scope. Defer.
Phase 2 file-share above 64KB exists in upstream; not in v0.1 OSS scope. Defer.
IETF SCITT compatibility deferred until at least one user explicitly requests it. Won't change pre-1.0.
Compatibility shims to standards-bodies are out-of-scope for v0.1 OSS. Existed in earlier R&D. Defer.
Install path requires only Python ≥3.10. No GitHub account required. Won't change.
Self-hosted relay defaults to plain HTTP on localhost; operator chooses ingress. Won't change.
Foreground-first. wire daemonize opt-in. Won't change.
wire is transport, not a memory backend. Agents bring their own memory (a vault, a DB, whatever). We won't ship a "structured cross-agent memory" API — the thing BAND (band.ai) gates behind Enterprise and leaves off its own real-time bus, absent from its own core model. A message bus that also owns your agents' memory is two products; we are one. Won't change. (Deliberate divergence — verified against BAND's shipped API, 2026-07-04.)
No per-room task board, no assignments, no goal tracker. Coordinating what work happens is the agent's job (or its orchestrator's); wire coordinates messages between identities. BAND ships a beta kanban; that's orchestration, a different lane (see #8). Won't change.
We won't build an approval-queue, a pre-action pause/kill console, or a policy engine as a product surface. wire's honest governance is the signed, verifiable audit trail it already emits (every envelope Ed25519-signed; see #11) — a stronger evidentiary claim than a self-reported event log, and it needs no dashboard to be true. Surfacing that audit queryably is fine; building "governance" theater on spec is not. Won't change pre-1.0.
OSS projects die from scope creep faster than from any other cause. This list is the maintainer's pre-commitment device. Feature requests that violate any item: closed with link to this file.
If your need conflicts — fork. AGPL/Apache/MIT licenses make forks trivial. The point of OSS is that you don't need our permission.