Integration Partners
Your customers want AI inside your product. Instead of building it from scratch, expose an API and let Pinchy be the AI surface. One extension, and your product is AI-native.
Your customers want an AI assistant inside your product. They want natural-language access to their data, multi-step workflows, inbox-type automation. Building that well takes a team, a year, and an ongoing cost you didn't plan for.
Meanwhile, your customers are quietly evaluating tools that can already do it — just not with your data, your domain, or your terminology.
There is a third option: you don't build the AI, and you don't lose the workflow. You expose an API and let your customers run Pinchy as their central AI tool — with a purpose-built extension for your product.
The Model
Your product becomes accessible to your customers' AI — without you owning the AI.
You already have an API, or you extend the one you have with the specific operations AI agents need. No Pinchy-specific dependency in your codebase.
A Pinchy plugin wraps your API and exposes it to the agent layer as governed tools: each agent gets an explicit admin-controlled allow-list (agents start with zero tools), access is scoped to shared vs personal agents with Admin/Member roles, and every action is recorded in a signed audit trail. We build this with you in the first projects.
Self-hosted on your own infrastructure, with your extension installed. Pinchy becomes the interface through which they use your product — natural language, cross-system, auditable.
Typical Scenarios
Your customers configure a complex product at setup time — shifts, tariffs, policies, integrations. A Pinchy agent ingests their legacy data (PDFs, Excel, Word), asks clarifying questions, validates constraints, and produces a working configuration. Weeks of onboarding shrink to days.
Your customers already have a support product; they need a natural-language assistant that speaks its language. A Pinchy agent with your product's extension plus your knowledge base drafts answers, cites sources, and routes escalations. Your product gets an AI surface without your team maintaining one.
These are the two patterns we've seen repeatedly. The mechanism is the same in both: an API on your side, an extension in Pinchy, and the agent experience your customers actually want.
How We Work Together
The extension mechanism works, but the public developer docs aren't mature enough to hand off unsupervised. For early partners, that means direct work with the core Pinchy team.
We map your product's domain to the Pinchy agent model. Which operations become tools. Which data becomes knowledge. Which steps need human approval.
We build the extension with you. Your team keeps ownership of the API. We take ownership of the Pinchy side, including permissions, audit, UI hooks.
The extension is packaged and shipped to your customers' Pinchy instances. Configuration, templates, and agent presets ship alongside.
Each completed integration sharpens the extension surface and the docs. The goal is self-serve — but only once the mechanism is proven across enough real projects to trust on its own.
If you build software and your customers are asking for AI, there's a good chance this model fits you. A 30-minute call is usually enough to tell whether it does.
30 minutes. We'll look at your product and tell you whether this fits — or what would need to change if it doesn't.
FAQ
Pinchy becomes the AI layer your customers interact with, while your product exposes its domain via an API. You get AI-native workflows — natural-language assistants, document analysis, multi-step agents with permissions and audit — without building or maintaining them yourself.
No. That's the point. Instead of every SaaS vendor bolting on their own AI feature set, you expose an API and we build a Pinchy extension that wraps it. Your customers get an AI-native experience; you keep your focus on your core product.
Not yet. The extension mechanism exists and works in real customer deployments, but the public developer documentation is not mature enough to hand off unsupervised. Early partners work with us directly; as the extension surface stabilizes across real projects, we'll publish self-serve docs.
Commercially, we work out terms per engagement. Technically, the extension is a thin adapter between your API and Pinchy's plugin model. Your domain logic stays yours; Pinchy provides the AI surface, permissions, audit, and UI.