Real problems, solved in production.

Challenge

A food brand's product data lives in too many places at once. The ingredient statement, nutrition panel, and allergens differ by pack size; the marketing site, Kroger's listings, and every distributor each hold their own copy; and product photography drifts out of sync across retailers. Change a recipe and someone has to chase that change through a dozen systems by hand, which is how a stale allergen line ends up in front of a shopper.

Solution

Kuhn Consulting designed and built ProvisionSync, a central platform where a brand manages its product data once and publishes it everywhere it needs to appear.

• One catalog for ingredients, nutrition, and allergens across every packaging format, with a live label preview as you edit
• Embeddable widgets (nutrition labels, a store locator, recipe cards, event calendars) that drop onto Squarespace, WordPress, or any site with one script tag and update the moment the data changes
• An Embed Designer that restyles those widgets to match the brand's own site, with an optional AI pass that reads the site and suggests on-brand looks
• Store locations and shelf availability synced from the Kroger API and Destini, deduped into one retail footprint with per-product stock and pricing
• A shared home for product images with guided distribution out to retailers and distributors, so imagery stays consistent across providers
• Multi-tenant, white-label architecture: one build serves every brand, each on its own domain and skin, with a read-only public API behind CORS and edge caching

Results

A production platform, live at provisionsync.com, where a brand edits a product once and every page that shows it follows. Label data, store availability, and imagery stay in step across the brand's own site and its retail partners, instead of being copied by hand and going stale.

Challenge

Small federal contractors compete against primes with dedicated capture teams. SAM.gov and Grants.gov post thousands of notices a week; figuring out which ones are worth bidding (the right NAICS codes, set-asides you actually qualify for, agencies you know) takes hours of scrolling, and the proposal work that follows starts from a blank page.

Solution

Kuhn Consulting designed and built Cairn Federal, a multi-tenant SaaS platform that guides small-business contractors from first notice to submitted proposal.

• Fit-ranked search: SAM.gov and Grants.gov notices scored 0 to 100 against NAICS codes, set-aside eligibility, agency history, and value band, with the reasons spelled out
• SAM.gov profile auto-fill from a UEI, including auto-detected set-aside certifications; USAspending award history for competitive intelligence
• Five-stage capture pipeline with bid/no-bid scorecards, gate reviews, and red-flagged deadlines
• AI capture assistant that acts on the pipeline (advances stages, drafts summaries, flags gaps), with every write gated behind user confirmation
• Solicitation shred: tears an RFP into a requirements checklist and compliance matrix
• Proposal drafting grounded in the company's own past proposals, so first drafts sound like the team wrote them
• Saved searches with scheduled email alerts for new matching notices

Results

A production platform, live at cairnfed.ai, where a contractor connects their SAM.gov profile in minutes and works from a ranked shortlist instead of a thousand open tabs. The AI handles summaries, shreds, and first drafts; every change to the pipeline stays under human control.

Challenge

AI assistants and agents are only as useful as the tools they can reach. But exposing an existing REST API to a model like Claude means hand-writing and maintaining a bespoke Model Context Protocol (MCP) server for every endpoint. That work is repetitive, error-prone, and out of date the moment the API changes.

Solution

Kuhn Consulting designed and built API2MCP, a platform that turns any REST API into a hosted MCP server, so AI clients can call existing services as native tools with no custom glue code.

• Register API providers from an OpenAPI spec or auto-import existing endpoints
• One-click “Wrap to MCP” generates a live MCP server and toolset
• Tools stay in sync with the live API spec: new endpoints added, removed ones dropped, and manual tools kept
• Secure credential and auth management per provider (API key, bearer header, and more)
• Centralized dashboard for servers, tools, providers, and 24-hour usage
• Tier-based plans with rate limiting, payload limits, and per-day request quotas

Results

A hosted platform where teams stand up production MCP servers in minutes instead of days, turning live APIs like weather.gov and USDA into AI-ready tools that Claude, Claude Code, and other MCP clients can call directly.

Challenge

Multiple client applications needed the same shared backend infrastructure: authentication, payments, AI integration, messaging. Building all of that separately for each app was costly and unsustainable.

Solution

Kuhn Consulting architected and built a multi-tenant SaaS backend serving as shared infrastructure, with open-source SDK packages enabling rapid client app development.

• Multi-tenant architecture with database-level tenant isolation
• Multi-provider AI integration layer (OpenAI, Azure OpenAI, Anthropic) with prompt versioning and streaming
• Stripe + PayPal subscription billing with full lifecycle management
• Enterprise auth: OAuth2, WebAuthn/Passkeys, 2FA, social login
• Multi-provider SMS/email services with templated messaging
• Shared TypeScript SDK packages for React, React Native, and Node.js
• Real-time messaging hub via SignalR
• Reusable Blazor component library with theming and centralized error handling

Results

A production platform now serving multiple client applications from a single backend, with reusable SDK packages that dramatically reduce development time for new apps. Full subscription management, real-time messaging, and AI capabilities available to all tenants.

Challenge

A client needed an AI coaching app that could run natively on every major platform (Windows, iOS, macOS, and Android) with real-time streaming, voice interaction, and the flexibility to switch between AI providers.

Solution

Kuhn Consulting built a cross-platform .NET MAUI application with Blazor Hybrid UI, delivering native performance from a single C# codebase.

• Multi-model support with provider abstraction (OpenAI, Anthropic, Gemini)
• Real-time streaming AI responses with persistent session history
• Speech-to-text and text-to-speech for hands-free voice coaching
• Configurable coaching style and model parameters
• Tier-based subscription with usage limits

Results

A fully native AI coaching app deployed across all four major platforms from a single codebase, with smooth voice interaction and the ability to switch AI providers without code changes.