Build, deploy, and sell AI agents for local-service businesses - from your IDE.
The MCP server enables building, deploying, and selling AI agents for local-service businesses directly from your IDE. It supports triggers, skills, channels, tools, and memory, with guardrails and deploy workflows to answer phones, chats, and SMS while managing real bookings.
SeldonFrame is the agent-native, open-source alternative to GoHighLevel for agencies serving local businesses. Give your coding agent a client's URL and it can create a branded website, booking flow, intake, CRM, and AI agent in one workspace. Run evals, publish, and hand it off β then operate every client from one agency view.
> Build a branded AI front office for my HVAC client in New Orleans.
β Workspace ready at acme-hvac.app.seldonframe.com
Build the first client workspace before paying. Hosted Managed uses SeldonFrame's keys; Builder and Agency plans are BYOK. Your agency owns the client relationship, workspace data, and delivery margin.
See a real one β a live workspace, not a mockup:
The generated site β quote intake, reviews, and the chat agent (bottom right)
Its booking page β the real calendar the agent books against
And here's the agent on that site handling a real inbound β checking the live calendar, and, when asked for a price, refusing to invent one (that's the quote-guard, not a scripted reply):
A genuine exchange with the chatbot on the live site β nothing scripted:
It reads the real calendar. "Today's fully booked β here are tomorrow's 9/10/11 AM slots."
It won't lie about price. Asked what a visit costs, it says "I don't have exact rates on hand, but I'll have a technician confirm the cost before any work starts" β the quote-guard in action, not a canned line.
It qualifies to book. Asks for the service address to confirm you're in-area, then locks the slot.
Same agent answers on SMS and voice too β the channel is just a config knob.
Start with the client URL. SeldonFrame extracts the business facts and creates the workspace.
Apply your agency playbook. Brand the site, booking flow, intake, CRM, and agent.
Verify before handoff. Run evals and complete a real test booking.
Publish and operate. Give the client a branded portal while your agency keeps the master view.
The same loop works from the dashboard, Claude Code, Cursor, Windsurf, or any MCP-compatible agent. See the agency workflow.
What an agent is made of here
Most "agent frameworks" hand you a chat loop and wish you luck. SeldonFrame gives your IDE agent the full production anatomy:
The agent model β Trigger Γ Skill Γ Channel
An agent isn't a chatbot UI; it's three independent axes:
Trigger β when it runs: inbound (a call / chat / email / SMS arrives) Β· event (a domain event fires β booking.completed, lead.created, invoice.paidβ¦) Β· schedule (a cron cadence).
Skill β what it does: receptionist Β· review-requester Β· speed-to-lead Β· win-back Β· digestβ¦
Channel β how it speaks: voice Β· web chat Β· SMS Β· email Β· internal digest.
Tools β what it can touch: native tools (book against the real calendar Β· read/write the CRM Β· send SMS & email Β· take a message) plus 1,000+ app integrations via Composio (Google Calendar, Sheets, Slack, HubSpot, Notion, β¦), bound per-agent.
surface: voice | chat (the old receptionist-only knob) is just one point in this space β trigger=inbound. One builder creates any agent; the marketplace sells any agent.
From agent to production loop
A production agent is a loop, not a single prompt:
Trigger β (Model + Tools + State) β Verify β Iterate, bounded by a Stop condition, improved by Evals, kept honest by Observability + Guardrails.
Two non-negotiables drive the roadmap: the checker must be separate from the maker (a model grading its own work is too generous a grader), and the loop must have brakes (or it bills you in silence). Where each primitive stands today:
Primitive
Status
What's there
Trigger
β Shipped
Inbound + event triggers on the SeldonEvent bus. booking.completed β review-requester; lead.created β speed-to-lead, both sending outbound SMS/email.
State
β Shipped
Agent loop-memory in Brain v2 β agents recall what they did before acting and record after. The review "ask once per customer" throttle is now a memory recall, not a bespoke flag.
Verify (maker β checker)
β Shipped
Deterministic validators grade every run β pass rates surface on each agent's health card and /runs; run_agent_evals replays scripted scenarios. Rolling out: the same checker as a hard pre-send gate + an LLM judge for judgment calls.
Guardrails / Stop
β Shipped
Quote-guard (never invent prices), enforced read-back before booking, per-contact throttles, booking-policy enforcement (hours Β· duration Β· required fields), hard call/iteration caps. Rolling out: generic token-budget brakes for long-looping agents.
Generate-by-default
β Shipped
One English sentence β trigger + skill + channel + tool bindings + guardrails, generated together. "text every customer for a Google review the day after their job β never twice, only if completed" compiles into a real agent. The generate_agent MCP tool and the Studio wizard both run this pipeline.
Record β Agent
πΊ Roadmap
Skip describing it β screen-share your workflow once. SeldonFrame watches, compiles the steps into a draft agent (skill + tool bindings + eval scenarios derived from the recording), and you deploy in minutes. Live behind a flag at /record; becoming the default on-ramp next.
The loop, drawn
One sentence builds it; triggers drive it; the ring is the loop running β Model β Tools β State β Verify β with tools orbiting the core, /runs + Brain observing, and the Stop brakes bounding every pass. Everything in the diagram is shipped today. (Animated SVG β if your viewer freezes it, the still frame reads the same.)
Deeper β the pre-wired stack, the architectural bet, the roadmap: docs/ARCHITECTURE.md.
Five agents you can ship today
Trigger Γ Skill Γ Channel are independent axes β pick one of each and you've architected an agent:
Agent
Trigger
Skill
Channel
The interesting part
AI receptionist
inbound call
qualify + book
voice
Books against the real calendar. Quote-guard, enforced read-back, take_message escalation to the owner.
Website chatbot
inbound chat
the same receptionist skill
web embed
Same blueprint, different channel β channels are swappable.
Speed-to-lead
lead.created event
first-touch + qualify
SMS
Fires seconds after a form submit, straight off the event bus.
Review requester
booking.completed event
review ask
SMS / email
Brain recall enforces "never ask the same customer twice."
Anything you describe
any
one English sentence
any
The Studio wizard adapts its questions to the primitives your agent needs.
The Brain β memory that compounds
Agents that forget are demos. Every SeldonFrame workspace owns a Brain:
The Soul β the single source of truth an agent grounds on (identity, services, hours, pricing). Install someone else's agent and it re-grounds on your Soul, never their facts.
Loop memory β agents recall before acting and record after. "Did I already ask this customer for a review?" is a memory recall, not a bespoke flag.
Notes β distilled rules β raw observations get compiled and promoted into durable knowledge on a schedule, so every run starts smarter than the last one ended.
Money rails, built in
Agents that do real work end up touching real money β so the rails are platform primitives, not an integration project:
A prepaid wallet ledger β UNIQUE idempotency keys and guarded never-negative decrements: a charge can't double-fire, a balance can't go below zero.
Stripe Connect payouts β seldonframe payout moves accrued earnings to a bank account.
Checkout that fulfills β a marketplace sale auto-provisions the deployment and re-grounds it on the buyer's business.
Per-use rental billing β every agent exposes a signed MCP endpoint; other LLMs can rent it and the ledger meters each use.
Real phone numbers β provision a number per agent (bring your own Twilio today; SF-managed metered numbers are rolling out).
Fee fine print: SeldonFrame takes a 5% GMV fee only when the marketplace brings the buyer, ~2% on sales through your own SeldonFrame storefront, and $0 anywhere else.
Operate every client from any IDE
One npm package β @seldonframe/mcp β runs as a local MCP server in every major AI-native editor. Pick yours, paste the snippet, and ask your agent to create or update a client workspace β website, booking page, intake form, CRM, and agent in one pass.
Claude Code
bash
claude mcp add seldonframe -- npx -y @seldonframe/mcp
Or as a plugin (auto-updates, one-liner via the marketplace):
Prefer to run it yourself? The entire monorepo is AGPL-3.0 β and there's a prebuilt image, so you don't even build it:
bash
git clone https://github.com/seldonframe/seldonframe.git && cd seldonframe
cp .env.docker.example .env.docker # add your ANTHROPIC_API_KEY or OPENAI_API_KEY
docker compose -f docker-compose.yml -f docker-compose.ghcr.yml up # pulls the prebuilt image
Brings up Postgres, migrations, and the app on localhost:3000 β the same dashboard, generated public sites, and API as the hosted version. Multi-arch image (amd64 + arm64) at ghcr.io/seldonframe/seldonframe; or docker compose up --build to build from source. Bring your own LLM key (Anthropic or OpenAI); SMS/voice (Twilio), email (Resend), and 1,000+ integrations (Composio) are add-your-own-key. Full guide: QUICKSTART.md.
When you're ready to scale
Your agency hub is seldonframe.com/agencies β deploy branded client workspaces, set your own client pricing, and keep the master view. Builders can still publish portable agents through seldonframe.com/build and let other LLMs rent them over MCP.
Pricing β no surprises
Self-host
$0 β AGPL-3.0, the entire monorepo Β· docker compose up and you're running
Builder
$29/mo Β· BYOK Β· workspaces for businesses you operate
Create the first client workspace free, then choose the hosted plan that matches your delivery needs
Marketplace sales
5% only when the marketplace brings the buyer
Tech stack
Next.js 16 (App Router) + React 19 + TypeScript β one deployable app: dashboard, generated public sites, and API
Drizzle ORM on Postgres (Neon in production; any Postgres 15+ when self-hosting)
Tailwind CSS 4
pnpm workspaces + Turborepo (monorepo)
MCP server β @seldonframe/mcp, plain Node, runs locally inside your IDE
Anthropic / OpenAI (bring your own key) Β· Stripe payments Β· Twilio voice/SMS Β· Resend email
Contributing
The highest-leverage PR here is an agent template or a vertical skill-pack β merged templates ship to the marketplace where every SeldonFrame user can find them, and you can list your own paid variants alongside. Core, connector, and eval PRs are equally welcome.
Start with CONTRIBUTING.md, then issues labeled good first issue / help wanted
House rule: agent-behavior changes ship with eval scenarios; runtime changes ship with tests
Community
π¬ Discord β fastest way to get help, feedback, or just say hi
AGPL-3.0 for the whole monorepo. Self-host freely; if you modify it and run it as a network service, your modifications stay open. For closed-source embedding, the hosted plan is the commercial alternative β see LICENSING.md. Same dual model as Mattermost, Plausible, and Postiz.
Build an agent. Sell it. Get paid. β from your IDE.
If this is the platform you've been looking for, β star the repo β it helps more builders find it.