Frequently Asked Questions

Common questions about Warburton, the Community Operating System for online communities.

What is Warburton?

Warburton is the Community Operating System for online communities, helping human moderators govern safer, healthier and more sustainable communities through AI-assisted governance, operational assurance and community memory.

It runs inside existing Telegram groups, Discord servers, and WhatsApp communities. Nine interlocking capability areas cover community health monitoring, operational assurance, protection protocols, knowledge capture, welfare escalation, governance, daily digest, deterministic safety, and contributor recognition — deployed as a single platform with a consistent character voice.

Is Warburton a moderation bot?

No. A moderation bot deletes messages and issues templated warnings. Warburton is an operating system for the entire lifecycle of community management.

Where a moderation bot gives you keyword filtering and ban commands, Warburton provides community health monitoring, decision provenance, deterministic protection protocols, knowledge capture with human attribution, welfare escalation, regulatory-grade governance, and deterministic safety rules. See Warburton vs Discord AutoMod for a detailed comparison.

How is Warburton different from Discord AutoMod?

Discord AutoMod provides keyword filtering, spam detection, and mention limits within Discord. Warburton operates across Telegram, Discord, and WhatsApp and provides nine capability areas that go far beyond content filtering — community health monitoring, full decision provenance, deterministic protection protocols, knowledge capture, welfare escalation, and regulatory-grade governance with hash-chained audit trails.

AutoMod is a content filter. Warburton is a platform. See the full comparison.

Does Warburton replace human moderators?

No. Warburton augments human moderators. It handles monitoring, detection, and routine responses while routing complex situations — welfare crises, ambiguous moderation calls, and edge cases — to human operators. Nine kill switches allow operators to disable any subsystem instantly. Crisis-level welfare events always route to a human. The Daily Digest gives moderators a structured summary so they stay informed without being online constantly.

Does Warburton use AI?

Yes. Warburton uses a tiered multi-model architecture: local Ollama models handle most requests, Granite provides automatic fallback, and Claude is escalated to for complex queries.

However, safety-critical rules are deterministic — enforced at the system level with no AI in the loop. Eight hard rules fire every time without exception, regardless of what any language model decides. The AI generates responses; the deterministic layer constrains them.

How does Warburton handle privacy?

All data is scoped per-user and per-tenant. Memory queries, learned cache responses, and audit records are isolated per deployment. Tenant-context wrappers are applied consistently across all three platform adapters. Web fetch is allowlist-only with SSRF defences. The dashboard is tenant-scoped with CSRF protection. Learned cache replies have usernames stripped.

The consent ledger uses FCA-aligned 2,562-day retention periods. Every consent event is recorded with a timestamp and is immutable once created.

What platforms does Warburton support?

Telegram, Discord, and WhatsApp. All three platforms are supported in every subscription at no extra charge. Warburton connects to existing communities — no migration required, no new platform to learn.

How much does Warburton cost?

$2,586 per month plus tax. A 14-day free pilot is included. No setup fees and no long-term contract required. Costwolds LLC configures Warburton for the client's community during the pilot period. Warburton is live and moderating within days of sign-up.

Who built Warburton?

Warburton is built and operated by Costwolds LLC.

How is Warburton's governance different from logging?

Most platforms log events but not decisions. They record what happened but not why. Warburton's governance layer records the complete decision chain — the policy that applied, the routing path that was taken, and the reason for the outcome.

Audit trails are hash-chained so tampering is detectable. Incident views are structured for direct board reporting without manual summarisation. The consent ledger uses FCA-aligned retention. This is governance built to a financial services standard, not an afterthought.

What can Warburton help answer?

Warburton helps community operators answer questions like:

• Is this discussion becoming unhealthy?
• Are moderators becoming overloaded?
• Has this topic been discussed before?
• Is somebody being unfairly targeted?
• Should this conversation be escalated?
• What happened overnight?
• What changed in community health this week?
• Why was this moderation action taken?
• Is this person showing signs of distress?
• Who contributed the answer to this question?