Ctrl-Fabric · Pitch deck v2
Governed APIs.
Governed agents.
One control plane.
APIFrames and Agent Orchestration — wrap APIs once, stand up AI agents safely, and keep every call inside the same guardrails.
Ctrl-FabricBrand
We build products, bespoke applications, and advisory for developers and enterprises — focused on stronger security and centralized control across your stack.
APIFrames
API wrappers, MCP servers & docs · apiframes.com
Wrap once. Govern every caller — humans, services, and AI agents.
runningbots
Agents & APIs · runningbots.ai
Orchestrate agents, wire them to governed APIs, audit every call.
Ctrl-FabricThe problem
AI is racing ahead of API governance.
Enterprises connect copilots and agents to production systems faster than they can secure, subset, and audit those connections.
API sprawl without a control plane
- Auth, quotas, throttling, and audit bolted onto every service
- Partners, teams, and agents need different slices of the same catalog
- Docs and agent tool lists drift from the real access surface
Agents without enterprise rails
- Per-integration glue for every LLM, tool, and knowledge source
- No least-privilege tool surface for copilots and swarms
- Hard to observe runs, cap usage, or revoke access instantly
Ctrl-FabricMarket opportunity
Two large markets converging on one control plane.
- [1] Expert Market Research — API Management Market (2025 base ~USD 5.84B; CAGR ~29.5% 2026–2035). expertmarketresearch.com
- [2] Market Research Future — API Management Software (~USD 18.21B in 2025; CAGR ~16.34% to 2035). marketresearchfuture.com
- [3] Technavio — API Management growth analysis 2025–2029 (CAGR ~12.3%). technavio.com
- [4] Precedence Research — API Management (~USD 12.16B in 2025; CAGR ~34% to 2034). precedenceresearch.com
- [5] Fortune Business Insights — Agentic AI Market. fortunebusinessinsights.com
- [6] The Business Research Company — AI Agents Global Market Report. thebusinessresearchcompany.com
Ctrl-FabricDemand signals
Adoption is accelerating — production governance is not.
The gap between “we plan to use agents” and “we can govern every tool call” is the buying trigger.
- [1] Gartner — “40% of Enterprise Apps Will Feature Task-Specific AI Agents by 2026…” (26 Aug 2025; updated 5 Sep 2025). gartner.com/newsroom
- [2] McKinsey / QuantumBlack — The State of AI 2025 (agentic scaling 23%; experimenting 39%). mckinsey.com
- [3] Anthropic — MCP ecosystem / donation to Linux Foundation Agentic AI Foundation (Dec 2025): 10,000+ public servers; 97M+ monthly SDK downloads. See Anthropic AAIF/MCP announcement.
Ctrl-FabricProblems we solve
Where projects stall — and where we intervene.
What breaks
- Pilot → production gap (McKinsey: scaling still narrow)[1]
- MCP tool sprawl outruns auth, allowlists, and audit[2][3]
- Full OpenAPI dumps into agents violate least privilege
- Split control planes: gateways for humans, frameworks for agents
- Governance / ROI / cost cited among top agentic program risks[4]
What Ctrl-Fabric collapses
- One wrapper: auth, quotas, throttle, audit
- API Views → lean MCP tools from one definition
- Config-driven agents/swarms with runs & traces
- Identity-backed tenancy, caps, kill-switch
- Humans, services, and agents on one governed edge
- [1] McKinsey State of AI 2025 — 23% scaling agentic somewhere; ≤10% per function. mckinsey.com
- [2] Stacklok 2026 software report (via Digital Applied) — ~41% of surveyed software orgs report limited/broad MCP production; security a top blocker. digitalapplied.com
- [3] Industry MCP gateway guidance (e.g. MintMCP) — enterprise need for centralized auth, tool allowlists, and audit. mintmcp.com
- [4] Secondary coverage of Gartner agentic project-risk commentary (governance, ROI, cost) — treat as directional; verify against latest Gartner note before diligence.
Ctrl-FabricIdeal customer profile Working ICP
Who buys — and who uses.
Primary buyer
- Platform / API engineering lead or Head of Developer Experience
- Owns API gateway, partner APIs, or internal platform
- Now asked to “make agents safe” without a new stack
Economic buyer
- CISO / Head of Security or VP Engineering
- Cares about least privilege, audit, revoke, blast radius
- Blocks ungoverned MCP and shadow agent tools
Champion / user
- Product + AI eng shipping copilots / ops agents
- Needs MCP + RAG + runs without a quarter of plumbing
- Use cases: deal-desk, compliance, ops, integrations
Derived from product positioning (landing + pitch). Not yet validated with closed-won deals — treat as hypothesis for discovery.
- [1] Ctrl-Fabric product positioning — ctrlfabric.com home & products config; APIFrames / runningbots landing copy (partners, teams, agents, deal-desk / compliance / ops).
- [2] Buyer / champion / anti-ICP framing is a working hypothesis for discovery — not closed-won ICP research.
Ctrl-FabricBottom-up market Illustrative model
TAM → SAM → SOM (working numbers).
Assumption ladder (replace with pipeline math)
Year 1: 8–15 paid pilots @ $24–48k · Year 2: 40–60 logos @ $48–72k · Year 3: 100–200 logos @ $60–120k as Enterprise + usage attach. Convert design partners via assisted onboarding (current runningbots model).
- [1–4] API management sizing — Expert Market Research, Market Research Future, Technavio, Precedence Research (see Market slide for links).
- [5] Fortune Business Insights — Agentic AI Market (~$7.29B in 2025). fortunebusinessinsights.com
- [6] The Business Research Company — AI Agents (~$8.29B in 2025). thebusinessresearchcompany.com
- [A] SAM/SOM and ACV bands are Ctrl-Fabric planning hypotheses, not third-party TAM research — recalibrate after discovery calls.
Ctrl-FabricSolution · Ctrl-Fabric products
Two products. One stack story.
APIFrames
Wrap APIs once. Control, audit & expose them everywhere — partners, teams, and AI agents — with curated subsets, branded docs, and MCP under the same guardrails.
Launch 01 · apiframes.com
runningbots
Configuration-driven agents and swarms — gateways, skills, MCP, knowledge bases, runs, traces — wired to governed APIs from one console.
Launch 02 · runningbots.ai
Wrap once. Orchestrate agents. Govern everywhere.
Ctrl-FabricProduct · APIFrames
Three pillars. One control plane.
Abstraction layer
Auth, usage caps, rate limits, and audit before traffic hits backends.
Curated subsets
API Views for partners, teams, mobile, agents — 403 at the wrapper for excluded ops.
Docs & MCP
Subset-scoped OpenAPI + MCP from the same definition — always in sync.
Ctrl-FabricProduct · Agent Orchestration
Stand up agents. Run them. Trace everything.
Config-driven agents
Personal / enterprise / global · micro · functional · router — skills, MCP, KBs without glue code.
Runtime & swarms
Chat, stream, async · bounded hierarchy · BYO OpenAI-compatible gateways · model budgets.
Observability
Runs, traces, tool calls, usage, latency, errors — agents operated like production services.
Ctrl-FabricPlatform
One console — agents & APIs.
Ctrl-FabricCompetitive landscape
Four lanes — we sit at the intersection.
API management
Kong · Apigee · AWS API GW · Azure APIM · MuleSoft · Tyk — strong HTTP policy; MCP/AI bolted on.
LLM / AI gateways
Portkey · LiteLLM · Kong AI — model routing & spend; not API subsetting or partner docs.
Agent frameworks / studios
LangGraph/LangSmith · CrewAI · Copilot Studio · Bedrock Agents · Agentforce — build/run agents; business-API governance often out of scope.
MCP gateways
Kong MCP · MintMCP-class — tool auth & allowlists; rarely paired with API Views + full agent orchestration.
Ctrl-FabricWhitespace
Governed subsets + agents + MCP, together.
| Capability | Classic APIM | LLM gateway | Agent framework | Ctrl-Fabric |
|---|---|---|---|---|
| Auth / rate / audit on business APIs | Strong | Limited | Limited | Core |
| Least-privilege subsets for partners & agents | Partial | No | No | API Views |
| Docs + MCP from same definition | Emerging | No | DIY | 1:1 sync |
| Config-driven agents / swarms / traces | No | Partial | Strong | Orchestration |
| Shared identity, caps, revoke | Partial | Spend | App-level | Platform |
Ctrl-FabricWin / loss themes Playbook
How deals are won — and lost.
vs Kong / Apigee / cloud APIM
- Win: API View → MCP in sync; agent runtime in-stack; faster subsetting for partners/agents
- Lose: ultra-high QPS mesh maturity, deep existing APIM investment, Gartner-box checklists
- Displace path: sit in front of upstreams for agent/partner slices; coexist with incumbent gateway
vs LangGraph / CrewAI / LangSmith
- Win: governed business APIs + Identity + ops console without assembling five tools
- Lose: deep graph/workflow DIY, OSS community, eval ecosystem
- Ally path: frameworks as skills; we own edge + tenancy + MCP publish
vs Copilot Studio / Agentforce
- Win: multi-LLM, API-first, not locked to M365 or Salesforce data plane
- Lose: if buyer wants “agents only inside their CRM/M365”
- Price foil: Studio credit packs (~$200/25k credits); Agentforce conversation/user SKUs — we sell control plane, not seat sprawl
vs MCP-only gateways
- Win: same subset drives docs + MCP + human API access
- Lose: if buyer only needs a thin MCP proxy tomorrow
- Upsell: start MCP publish from Views; expand to orchestration
Ctrl-FabricPricing & packaging Proposed — not live
Land with APIFrames. Expand to the platform.
Pilot
- 1 tenant · assisted onboarding
- Limited API Views / MCP servers
- 90-day convert-to-Team path
- Fits design-partner program
Team
- APIFrames: Views, MCP, docs, audit
- Call & MCP quotas (fair-use bands)
- Optional Agent Orchestration add-on
- Target ACV ~$48–96k
Enterprise / Platform
- Full runningbots console (agents + APIs)
- SSO, multi-region, RPA, SLAs
- Usage attach (calls, agents, seats)
- Target ACV $100k–250k+ (vs Kong/Apigee bands)
Today Identity plans are $0 (launch plumbing). Commercial SKUs not seeded. Proposed Pilot/Team/Enterprise bands are hypotheses for partner pricing talks.
- [1] Vendr marketplace summaries — Kong mid-market/enterprise ACV commonly ~$40k–$180k+ (deal-dependent). vendr.com/marketplace/kong
- [2] Google Cloud Apigee pricing (consumption + environment SKUs; enterprise deals often six figures). cloud.google.com/apigee/pricing
- [3] Microsoft Copilot Studio — Copilot Credit pack ~$200 / 25,000 credits (billed annually). learn.microsoft.com
- [4] Ctrl-Fabric proposed Pilot $2–4k/mo · Team $4–8k/mo · Enterprise custom — not live list prices.
Ctrl-FabricTraction & stage
Incubating in the open — shipping toward first release.
Where we are
- Stage: incubation · products marked coming soon on ctrlfabric.com
- Launch order: APIFrames first, then runningbots
- Product surface: live consoles & services in development (first-release focus on APIFrames)
- GTM motion: assisted tenant onboarding · early-access email CTAs
Proof points to fill
- Named design partners / pilot logos — none published yet
- Paid ACV / waitlist count — to instrument
- Public case study (partner subset + MCP + agent) — target for Launch 1
- Contact: support@ctrlfabric.com · “early access” / “request onboarding”
Honest stage slide: no fabricated logos or revenue. Replace this panel as pilots close.
Ctrl-FabricTeam
Building the control plane for agentic software.
Ctrl-Fabric
- Products, bespoke applications, and advisory focused on stronger security and centralized control
- Based in Bengaluru, India
- Domains: ctrlfabric.com · apiframes.com · runningbots.ai
- Contact: support@ctrlfabric.com · +91 89392 99806
Leadership
- Sridhar Prabhakar — Identity / platform author · LinkedIn
- Team strengths called out publicly: architecture, security & access, reliability, UX, agentic workflows
- Add full roster, advisors, and prior exits for investor versions
Ctrl-FabricThe ask
Design partners now. Capital when it accelerates the same plan.
Primary — design partners
- 8–15 pilot orgs over the next 2–3 quarters
- Commit: 1 governed API surface + 1 agent or MCP workload
- We provide: assisted onboarding, weekly office hours, roadmap influence
- You provide: production-shaped traffic, security review feedback, reference rights
- CTA: support@ctrlfabric.com · subject “APIFrames / runningbots early access”
Optional — fundraising Fill amount
- Ask: $___ (seed / pre-seed) — insert when ready
- Use of funds: Launch-1 hardening & hosting · 2–3 GTM/solutions hires · design-partner success · security/compliance packaging
- Milestones: GA APIFrames · 10 paying logos · runningbots GA · $1M ARR path clarity
- Not raising publicly until design-partner pipeline is real — avoid vanity round
Ctrl-FabricWhy now
MCP made the API edge the product.
- Major model vendors standardized on MCP for tools[1]
- ~41% of surveyed software orgs report MCP in limited/broad production[2]
- Security remains the #1 enterprise MCP blocker[2][3]
- Classic gateways see HTTP — not agent tool intent[3]
- Buyers need gateway + subsets + agent runtime, not three more point tools
- “Shadow MCP” must land under one inventory and policy[3]
- [1] Anthropic MCP ecosystem / AAIF (Dec 2025) — broad client adoption across major AI products; 10k+ public servers. See Anthropic AAIF/MCP announcement.
- [2] Stacklok 2026 software report summarized by Digital Applied — ~41% limited/broad MCP production; security a leading blocker. digitalapplied.com
- [3] MintMCP — enterprise MCP security checklist (gateway for auth, allowlists, audit). mintmcp.com
Ctrl-Fabric
Ctrl-Fabric
Govern the agentic stack.
Start at the API edge.
APIFrames first. runningbots next. One control plane for humans, services, and agents.
Ctrl-Fabric