CallHub is the unified outreach platform for 200K+ political campaigns, unions, nonprofits, and advocacies — calls, SMS, voice broadcasts, email, P2P texting, all coordinated by the Unified Goals Dashboard. callhub.io is already on Cloudflare DNS. The expansion footprint is the developer platform underneath: Workers + Queues for the multi-channel orchestration engine, AI Gateway for Call Sentiment Analysis + Agent Quality & Coaching, R2 for call recording + analytics archive, and Workers for Platforms for per-organization tenancy across unions, political, and nonprofit verticals.
callhub.io on Cloudflare via cloe + frank.ns.cloudflare.com, with server: cloudflare on every responsegaglers.com is also Cloudflare-fronted (cf-ray on every response) on Kinsta origin200K+ teams running calls, SMS, voice, email — across 8 channels with two-way CRM sync, AI-powered sentiment analysis, and agent coaching. The orchestration is your IP. The runtime underneath — telephony abuse defense, multi-channel orchestration state, AI inference for sentiment, and per-org tenancy across unions / political / nonprofit verticals — is what Cloudflare's developer platform was built to be.
Each maps to something published on callhub.io today — Call Sentiment Analysis, Agent Quality & Coaching, Unified Goals Dashboard, the 200+ integrations, the 10DLC SMS compliance work. Status tags show what's already on Cloudflare.
Authoritative DNS via cloe + frank.ns.cloudflare.com. Edge serves with server: cloudflare. The procurement is in place, SOC 2 mapping exists, MSA already negotiated.
"CallHub scores every call as positive, negative, or neutral while the campaign is still running." That's an LLM call per call. AI Mock Interviewer for agents is another. OpenAI is verified on your apex. AI Gateway gives you semantic cache, per-org attribution, BYO keys, and budget caps in one header change.
"Auto-rotate people into a phonebank when they engage with your text" is exactly a Workers + Queues pattern. Each engagement event triggers the next-best-channel decision. Workers run that decision at the POP closest to the supporter; Queues handle the durable fan-out into the right channel — no Kafka, no per-region SQS.
200,000+ campaign-organizing teams. Each one has its own 10DLC compliance posture, its own dialer numbers, its own CRM sync, its own GDPR/CCPA jurisdiction. Workers for Platforms gives each customer its own Worker namespace with isolated keys, egress, audit log — on the same edge.
Every call is recorded for QA + coaching + compliance. At 200K+ teams running multi-week campaigns, that's a serious storage + egress bill. R2 zero egress for replay, transcript pull, agent coaching review, regulatory retention. Plus Vectorize for "find me every call where a voter said X" semantic search.
Each live agent has session state: which voter they're on, the call buffer, the dispositions assigned, the post-call survey. Durable Objects give you a single-writer state holder per active agent session at the edge with native WebSocket — no Redis cluster, no sticky session router.
Free-tier outreach platforms attract scrapers, fake campaign signups, and unverified-sender exploitation. Bot Management at the edge stops the abuse before it reaches the 10DLC registration backend. Turnstile drops in on signup, sender verification, and supporter import flows.
NationBuilder, NGP VAN, PDI, Action Network, Salesforce, Blackbaud Luminate — 8+ named CRMs with two-way sync. Workers + Hyperdrive give you pooled, edge-resident connections to each CRM without sticky auth overhead — sync events flow through the same edge that orchestrates the campaign.
"Phonebanking and peer-to-peer texting" mobile app on iOS + Android. Volunteer state, dial-list cursors, supporter context, post-call disposition — all need low-latency global reads. KV + D1 give you that without standing up regional Postgres.
Two cost lines dominate as the platform scales: AI inference for Sentiment + Agent Coaching (one call per LLM, multiplied by every conversation) and telephony recording storage (regulatory-retained, multi-year). AI Gateway turns inference into a per-org attributable line; R2 turns recordings into a flat-rate archive.
Political campaigns spike at GOTV. Unions surge at contract negotiation deadlines. Nonprofits run year-round but spike at year-end. Churches need different content moderation. Each vertical has its own dialer pacing, its own 10DLC compliance posture, its own scale curve.
Every row sourced from public DNS, the callhub.io apex TXT, and HTTP response headers. The mint row is already live. The orange column is the additive overlay.
2026 is a major election year. GOTV in October/November will be the highest-volume use of CallHub the platform has ever seen. The architecture that handles that surge — multi-channel orchestration, sentiment at scale, agent coaching at scale, per-campaign tenancy — is the architecture that defines the next 4-year cycle. Picking the runtime now is materially cheaper than mid-cycle.
You're already on Cloudflare. cloe + frank.ns.cloudflare.com on callhub.io DNS today. Expanding from DNS + edge to AI Gateway + R2 + Workers for Platforms is a roadmap conversation, not a procurement event. The MSA is already in place.
OpenAI is already a vendor. Sentiment Analysis and AI Mock Interview both run through it. AI Gateway is the lowest-friction observability + cost-control upgrade available — no model migration, no prompt rewrite. Per-org attribution from request one is what makes Scale Plan pricing defensible to a CFO at a national union or major campaign.
The interesting conversation is which of these is closest to your current sprint: AI Gateway behind OpenAI, Workers + Queues for the Unified Goals Dashboard, R2 for call recording archive, or Workers for Platforms for the per-org tenancy story. I'd rather hear what's actually on your roadmap than guess.