ComingUp ComingUp
M

Multi-Tenant AI Agent Platform

Turn documented procedures into AI agents with live API endpoints in minutes

Jul 30, 2026 AI & Machine Learning
ai_agents multi-tenant saas source code white_label

About

This is a NON-EXCLUSIVE white-label LICENSE of the source code , you deploy under your brand and keep your revenue; I retain the IP. Full details below.License the full source of a production AI-agent SaaS , the entire lifecycle in one codebase: three ways to build agents (guided wizard, direct editor, and describe-it-in-plain-English — even from a screenshot), document knowledge with cited answers, 38 pluggable capabilities including governed email send, scheduled runs, Slack/webhook/database connectors, multiple AI models per workspace (Azure OpenAI, Anthropic, Google, OpenAI — bring your own keys), spend ceilings on every run, five-tier plans with live Stripe billing, per-step audit trails with one-click evidence-pack PDF export, and an embeddable anonymous public demo. ~74,000 lines. Deploy under YOUR brand; keep 100% of your revenue.What you're licensingThe complete source of a production AI-agent platform , not a single feature, the whole lifecycle:Build agents three ways: a guided wizard, a direct step editor, or Express mode — describe the goal in plain English (or drop in a screenshot of a form or a full SOP document) and the platform drafts the entire agent: instructions, step plan, required capabilities, and an honest list of gaps. Nothing publishes without human review.Ground them in knowledge: upload documents; agents answer from them with citations — and say so when the document doesn't cover it.Run them every way that matters: interactive chat workbench, real REST endpoints with scoped API keys (sync or async), and scheduled runs on cron-style timers.Give them capabilities: 38 pluggable tools , OCR and PDF extraction/generation, Slack posting, webhooks, read-only database query, text redaction, templated LLM steps, and a governed email-send tool (fail-closed kill switch, recipient caps, domain allow-lists).Choose the model per workspace: Azure OpenAI by default; Anthropic, Google, or OpenAI with the customer's own keys, behind one provider abstraction.Govern all of it: monthly spend ceilings enforced on every run type before any model call, mandatory review-before-publish, full per-step audit trails with tokens and cost, and one-click evidence-pack PDF export with a SHA-256 content digest.Monetize it: five-tier plan enforcement from a single rules table, self-serve signup (feature-flagged), and live Stripe billing where the dashboard IS the catalog — plus an embeddable anonymous public demo with its own spend circuit-breaker as your top of funnel.Built over 8 months by a 20+ year enterprise architect/CTO, running in production today with live Stripe billing. I run my own SaaS on this codebase — you get the same code, deploy it under your brand, and keep everything you earn. Non-exclusive: other licensees exist or will; what you're buying is a two-year engineering head start, not a monopoly.Why the architecture is worth licensingThe AI is ~20% of this codebase. The other 80% is what makes an AI product safe to hand to strangers — the part that takes a team a year: real multi-tenancy with isolation enforced at the data layer via global query filters; spend governance checked before any model call on every run type; plan enforcement that gates creation and never breaks a running workspace; Stripe as the catalog (add or reprice tiers with zero deploys); audit trails on everything; governed outbound email; an anonymous demo subsystem with a per-visitor limiter and a global daily circuit-breaker.Tech stackSingle .NET solution (C#, .NET 10) · EF Core · SQL Server / Azure SQL · server-rendered Razor (no SPA build chain) · Azure OpenAI + provider abstraction · Stripe · Docker (runs anywhere — proven outside Azure) · CI green, ~70% line coverage on the application layerBy the numbers~74,000 lines of application code (~85K with localization + docs) · ~430 C# files · 23 agent templates · 38 registered agent capabilities · 61-test unit suite + integration suite · full deployment runbooksLicense terms (plain-English summary — formal agreement at purchase)Non-exclusive, perpetual, worldwideDeploy and operate under your own brand, for your business and your clients (SaaS or client installs)Full source access; modify freelyNOT permitted: reselling, sublicensing, or redistributing the source itself — you sell the SERVICE it powers, not the codeYou receive the current release as a snapshot; 12 months of updates included, renewal optional afterward2 weeks of reasonable handover support (deployment + architecture walkthrough) includedTrademark/domains/brand not included (you rebrand); my patent filings not includedWho this is forA .NET agency or consultancy entering the AI-agent market with a two-year head start; an MSP deploying governed agents for clients under its own brand; a vertical-SaaS team that needs a proven multi-tenant AI execution layer. If you can read C# and run Docker, you can be white-labeled in a week.Why license instead of buildBecause I already paid the tuition: the cap-enforcement holes, the multi-tenant traps, the billing edge cases, the workbench-vs-production divergence — found, fixed, and tested. You start where I finished. Serious inquiries get the live demo, a video walkthrough, and the full diligence package under NDA , message me with what you'd deploy it for.

Comments (0)

No comments yet. Be the first to comment!