Replace fragile, custom AI integrations with a standardized, governed connectivity layer. One workflow live, 3–5 reusable MCP Servers, audit trails your security team can sign off on, and runbooks so your team operates it without us — fixed-price, beta from $4,500.
Your team writes one-off connectors for every AI-to-SaaS integration. They break every time Salesforce, Jira, or Workday pushes an update. Senior engineers spend cycles being plumbers.
Leadership can't answer 'what data did the agent touch, and under what permissions?' Without that answer, the project can't pass security review — and won't go live.
Because the official path is too slow, developers use unapproved tools and personal accounts. You know it's happening. Policy alone won't stop it.
Your 10–20 year old systems weren't built for AI. Forcing direct connections to a modern agent yields latency, hallucination, or accidental data exposure.
You're asked to show AI ROI this fiscal year — but most of the AI budget has been spent on sandbox pilots that never reached production.
You don't need more custom code. You need a contract between your agents and your systems. MCP is that contract.
MCP is the 'USB-C for AI.' One protocol, every system, no per-tool glue code.
We wrap your decade-old systems in a modern interface — no rewrite required.
Authentication, authorization, and audit are designed in on day one — not bolted on later.
Swap LLM providers or add a new MCP Server in days, not months.
The pilot workflow is selected on business impact, not 'cool factor.'
The same operating model on every engagement — what changes is the count of MCP Servers and the depth of governance.
Single job-to-be-done with the strongest measurable outcome. Selected together in a scoping session.
Read/write boundaries, auth approach, action constraints. Each system maps to one MCP Server.
Standardized servers for the agreed systems (the Bridge Factory). Tool schemas, action constraints, regression test harness.
Least-privilege access, scoped tokens, policy checks, allow-lists. OAuth / OIDC / API keys — SAML SSO is a separate engagement.
Activity logs, trace IDs, audit events per tool/action, cost budgets. The artifact your security team needs to sign off.
Test cases, prompt-injection considerations, break-glass controls, kill-switch.
Runbooks, ownership map, change process, walkthrough session. Your team owns it from day one.
Beta pricing is locked for the first 5 logos. After that, pricing moves to standard rates and onboarding becomes cohort-based.
14-day proof. One workflow. One MCP Server.
Production-grade. Most enterprise teams start here.
Multi-workflow rollout with custom policy.
If you complete the onboarding inputs on time (system access, documentation, SME availability), we guarantee delivery of the agreed artifacts by the end of the sprint: the working pilot workflow, the agreed MCP Servers, the governance controls, and the audit/ops handoff pack.
If we miss the agreed deliverables, you choose between a free extension week or a partial refund of fees attributable to the missed deliverables — capped at fees paid for the affected deliverables, per the MSA Addendum §11.
What we don't guarantee: business outcomes, regulatory compliance, security certification, or that the system is "secure by default." The pilot reduces risk through controls, auditability, and least-privilege design — but compliance determinations remain your responsibility.
The wrong answer is "yes, sign anyway." Read both columns before booking the call.
If your team spends even two senior engineers × 8 weeks on fragile integration plumbing, you've already spent more than the Standard Pilot — before auditability, security review, or ongoing break-fix.
| Option | Cost | Timeline | Catch |
|---|---|---|---|
| Hire in-house architect | $200–350K / year + benefits | 3–6 months to hire + 90 days ramp | Slow to start |
| Big-4 consulting firm | $500K – $2M+ | 9–24 months | Tech changes twice mid-flight |
| Internal team custom build | $0 cash · $150K+ opportunity cost | 4–6 months for v1 | Maintenance tax forever |
| Low-code wrapper SaaS | $500 – $5K / month + setup | Weeks | Vendor lock-in, weak on legacy / air-gap |
| HAPI MCP Pilot Accelerator | $4.5K – $50K one-time | 14 days – 12 weeks | Pilot scope only — rollout is separate |
Engagement mechanics, scope changes, and the questions security and procurement always ask.
The Thin-Slice covers one workflow and one MCP Server in a dev/test environment — enough to prove the pattern to your team and secure internal budget for the bigger engagement. The Standard Pilot covers one workflow but 3–5 MCP Servers, with full governance, audit, and a path toward staging. Many enterprise buyers run the Thin-Slice first as a low-risk evaluation.
That's the expected case. The engagement includes a walkthrough workshop, runbooks, and the 3 × 30-minute follow-up sessions over 3 months. Your team also gets free access to Rebel-U content during the engagement to ramp on MCP fundamentals while we build.
Your team owns and operates the system. You can either run it self-serve, or move to an optional Support Retainer ($2,500/month or $7,500/quarter) for ongoing MCP Server additions, governance evolution, and oncall office hours. Many clients also pick up a separate engagement to expand to more workflows after Q1.
Yes — extensions are handled via a change order to the original SOW. Common reasons: scope grew (extra MCP Server requested), a new IdP added, or you decided to push to staging earlier. Pricing is per-server or per-week depending on the scope of the change.
SAML / Enterprise SSO is a separate, dedicated SOW scoped per IdP (Okta, Entra ID, Ping, Keycloak). It's intentionally not bundled — IdP integrations are sensitive and require security-team coordination. We can sequence it after the Pilot or run in parallel with a separate signer on your side.
OpenAPI v3 is the fastest path because HAPI MCP can lift the spec directly into MCP tools. For non-OpenAPI systems (gRPC, internal RPC, raw databases, legacy SOAP), we build the MCP Server by hand — usually with a thin OpenAPI shim added during the engagement so the same standardized contract applies. This is in-scope; expect 1.5–2× the build time per non-OpenAPI server.
Fixed-price by default. The Pilot tiers are fixed-fee engagements with defined deliverables, a Ship-the-Foundation Guarantee, and predictable scope. T&M is only used for the Support Retainer and post-Pilot expansion work, where ongoing scope is harder to predict.
If you provide the agreed inputs (system access, documentation, SME availability) on time and we miss a Pilot deliverable, you choose between a free extension week or a partial refund of fees attributable to the missed deliverable. Refund is capped at fees paid for affected deliverables, per the MSA Addendum §11. This is your sole remedy under the Ship-the-Foundation Guarantee.
Beta pricing is capped at the first 5 logos and capacity-capped at 3 active pilots per month. Beta clients keep beta pricing for 12 months on follow-on MCP Servers added through a retainer or expansion SOW. After the first 5 logos, pricing moves to standard rates and onboarding becomes cohort-based.
The conversation takes 20 minutes. Either we agree it's a fit and you have a clear path to a signed SOW within a week — or it isn't, and you walk away with an honest answer and a free pointer to the free HAPI MCP version if that's a better fit for your team. No hard feelings either way.
Ship-the-Foundation Guarantee · Beta pricing locked for the first 5 logos · 3 active pilots per month.