Skip to content

Latest commit

 

History

History
307 lines (154 loc) · 8.23 KB

decocms.md

File metadata and controls

307 lines (154 loc) · 8.23 KB

DecoCMS.com  --- Positioning & Messaging

  1. One-Sentence Positioning

Headline (one-liner)

Deco is the home for all your AI apps.

Clarifier (immediately under it)

A generative control plane for data, tools, models, permissions, and memory --- so you can prototype in chat, ship in code, and run & monetize agentic apps in production.

Alt (analogy-first)

"As easy to start as a chat app, as programmable as a workflow engine---inside one governed admin."

  1. Why Now

  • Prompt → production gap: Non-technical teams can prototype, but production still needs policy, tenancy, logging, and cost governance.

  • Context is the product: Winners standardize on a governed context gateway (connectors + policy + observability) that agents and UIs share.

  • "AI app builders" ≠ ops: Many tools stop at pretty UIs. Companies need a runtime + control plane that's auditable, scalable, and extensible.

  • Monetization missing: The long tail of internal agents needs distribution + per-run billing to sustain an ecosystem.

  1. Category & Anchors

  • Category: Context CMS (open-source admin for AI agents)

  • Problem: Prototypes are easy; production (governance, costs, observability) is hard

  • Use case: Build, run, and monetize bespoke agentic apps on your own stack

  • Alternative: Stitching model APIs + workflow tools + SaaS + DIY governance

  • Persona: Web/TS devs & partner implementers; product/ops leaders

  1. Problem (tight)

Shipping production agentic software is messy. Teams glue together auth, data access, model routing, logging, costs, and approvals across vendors. Prototypes stall or get buried under compliance.

Sub-problems

  • Fragmented stack: LLMs here, workflows there, UI elsewhere

  • Governance/RBAC/FinOps bolted on late

  • "Vibe debugging" instead of repeatable ops

  • No clean way to reuse or monetize components

  1. Solution (what Deco is)

Deco is the Context Management System --- a control plane + runtime --- that unifies your agentic stack around your context.

  • Context Mesh (MCP-native): connectors to data/tools/models with policy, auth, RBAC, audit, spend caps

  • Unified runtime: compose agents, workflows, UIs in one repo; edge deploy; typed interfaces

  • Dual surface: prototype in chat, then harden/scale in code (SDK/CLI) without changing platforms

  • Observability: logs, traces, error analytics, cost per step

  • Marketplace + wallet: install or publish modules with one-click, per-run billing

  • Packages: full "OSs" can live inside DecoCMS (e.g., Storefront Suite, former deco.cx)

  1. Differentiation (how we're not "just another AI web app builder")

  • Control plane for context (not only code generation): governed routing, policy, and observability

  • Safe vibe coding: every run traced, every permission scoped, every cost capped

  • One codebase, one deploy: agents/workflows/UIs together; no vendor sprawl

  • Usage-based billing & monetization: build once, bill per run; transparent wallet billing

  • Open-source & extensible: self-host/BYOK; forkable core; partner ecosystem

  1. Promises (trust commitments)

  • Production-safe by default: RBAC, audit trails, approvals, spend caps

  • Portable & open: zero lock-in; your code, your data, your keys

  • Prompt-to-production: chat prototypes become shippable software---same runtime

  • Marketplace leverage: reuse what works; monetize what you create

  1. Capability → Feature → Benefit

  • Context Mesh → connectors + policy + RBAC → Right context at the right time---safely

  • Unified runtime → typed RPC + edge deploy → Prototype and ship on the same stack---no rewrites

  • Observability → logs & traces + cost per step → Operate with confidence; fix fast

  • Governance → audit, approvals, spend caps → Pass compliance; control risk & cost

  • Marketplace & wallet → one-click install; per-run billing → Reuse & monetize modules instantly

  • Packages → e.g., Storefront Suite → End-to-end "OSs" inside the same admin

  1. Single Story per Page (don't blend)

Story A --- Developers & Partners (/developers, GitHub)

H1: Open-source runtime for AI apps.

Sub: Build agents, workflows, and UIs in one repo; connect your stack; deploy to the edge; publish modules---without vendor sprawl.

Story B --- Teams & Founders (/teams or /founders)

H1: A single control plane for all your AI projects.

Sub: Prototype in chat, then hand off to engineers (yours or ours) to harden for production with roles, limits, audit, and clear costs.

  1. Homepage Copy (clear version)

H1: The admin for your AI agents.

H2: Connect context → compose agents, workflows, and UIs → deploy, observe, and bill per run.

Bullets

  • Prototype in chat; ship in code (same runtime)

  • Permissions, audit, spend caps out of the box

  • Install/publish via Marketplace; wallet billing

  • Open-source; self-host or BYOK

  • Train your team or tap certified partners

Primary CTA: Start with $5 credit (no credit card needed)

Secondary: See templates - Star on GitHub

  1. Elevator Pitches

General (45--60s)

AI made demos easy; production is where it breaks---permissions, logging, costs, and compliance are missing. DecoCMS is the admin for your AI agents: connect your context (data, tools, models, permissions, memory), compose agents/workflows/UIs, and run them in production with roles, audit, limits, and per-run billing. Prototype in chat, ship in code, and keep everything inside one governed admin. In 30 days, your first autonomous workflow goes live.

Investor-friendly (no jargon, 45--60s)

Next-gen companies won't win by renting bloated tools; they'll win by running software tailored to how they operate. DecoCMS turns a company's context into working automation---AI teammates and workflows that actually do the job. It's production-safe and extensible, with a marketplace so teams can install or sell modules with clear per-run billing. Customers ship a live workflow in a month and scale on their own terms.

  1. Sales Deck --- Titles & Bullets

Cover

DecoCMS --- the admin for your AI agents

  1. AI is easy to demo, hard to run

  • Prototypes multiply; production stalls

  • Permissions, audit, and cost control arrive too late

  1. You don't need more tools---you need a home

  • A control plane for context to operate agents

  • Apps that fit your processes and evolve with your data

  1. Meet DecoCMS (Context CMS)

  • MCP-native connectors + guardrails; unified runtime; edge deploy

  • Logs, tracing, No Vibe Debugging, cost per step

  1. Dual surface: chat for speed, SDK for scale
  • Business prototypes quickly; engineers productionize without switching stacks
  1. Packages & examples

  • Storefront Suite (ex-deco.cx) as an installable package

  • Templates: RFP Assistant, Fraud Triage, Content Ops, Scheduling, Knowledge Ops

  1. Marketplace & wallet

  • Build once, bill per run; one-click installs

  • Partners monetize repeatable solutions

  1. Open-source, zero lock-in
  • Self-host/BYOK; forkable core; composable components
  1. Next steps
  • Pick a template; start a pilot; certify your team
  1. Use-Case Templates (cross-industry)

  • Product Intake / Knowledge Ops (catalog & enrichment)

  • RFP Assistant (B2B sales ops)

  • Fraud Triage (fintech/transactions)

  • Content Ops (media/marketing)

  • Scheduling & Dispatch (field/logistics)

  • Onboarding & Training (HR/compliance)

Use retail examples in case studies without narrowing the headline category.

  1. Competitive Posture

  • Versus "AI app builders": DecoCMS includes the runtime + control plane, not just UI generation.

  • Versus workflow tools: We add context governance (RBAC, audit, spend caps) and typed, extensible code.

  • Versus vertical SaaS: Tailor-made apps on your stack; open-source, no lock-in; monetize your modules.

  1. Definitions

  • Context = data, tools, models, permissions, goals, memory

  • Context Mesh = governed connectors + routing + observability

  • Agentic app = automation that plans/acts with approvals and guardrails

  • Module = packaged agent/workflow/UI/connector installable via Marketplace

  • Per-run billing = wallet-based charging on each module execution

  1. Phrases to standardize

  • "Control plane for your AI stack."

  • "Prototype in chat. Ship in code."

  • "Safe Vibe Coding"

  • "Build once, bill per run."

  • "Own your context. Own your tools."