Day 38: v0.3.0
Released v0.3.0 today. And then spent the afternoon in a meeting that showed me exactly what Pinchy needs to become.
The Deployment Release
v0.3.0 is the release that makes Pinchy deployable by people who aren't me. VPS deployment guides with cloud-init automation, a loading page for first boot, HTTP compatibility for initial setup, and all the hardening from the past week. Plus the usage dashboard, so you can see what your agents cost from day one.
Also in this release: multi-bot Telegram support. Each agent can have its own Telegram bot, with permission-aware access control. If a user doesn't have access to an agent's group, the bot won't respond. This took days of edge-case fixing — pairing, unlinking, re-pairing, config conflicts — but it works now.
The Meeting That Wrote the Roadmap
Met with a recruitment company in Vienna. Four people around a table, laptops open, me demoing Pinchy live. Within ten minutes, the conversation shifted from "what can it do" to "here's what we need it to do."
Their world: freelancer invoices coming in via email, manual data entry into spreadsheets, contract generation from templates, billing reconciliation across multiple systems. Four or five steps that are all manual, all repetitive, all error-prone.
What they want: an agent that reads incoming invoices, extracts the data, matches freelancers by name (handling typos and variants), populates the spreadsheet, drafts the outgoing invoice, and creates an email draft for review. Not fully autonomous — human-in-the-loop for the final send. But everything up to that point: automated.
The conversation surfaced a pattern I keep seeing: companies don't need one magic agent. They need Pinchy connected to two or three systems they already use. For this company, that's Outlook and OneDrive. For the Odoo company, it's the ERP. For others, it's Jira or Slack or SharePoint.
The Corporate Perspective
Earlier in the day, a very different conversation: a corporate startup working on communication between humans, agents, and sub-agents. The kind of challenges that large enterprises face — where it's not just one company deploying agents, but agents from different organizations needing to communicate with each other.
Two meetings in one day, two completely different scales. A recruitment company that needs email automation and invoice processing. A corporate innovation team thinking about inter-company agent communication. Both want Pinchy, for very different reasons.
This is the question I keep circling: is Pinchy for SMEs or for enterprises? The honest answer right now is: I don't know yet. I'm deliberately staying open, talking to everyone, and watching where the pull is strongest. What I do know is that the core needs to stay compact and flexible — a solid foundation that works for a 5-person recruitment firm and a 50,000-person corporation, with the difference being in plugins, integrations, and configuration, not in the core product.
What This Means for the Product
The immediate roadmap is now crystal clear:
- Email integration — Outlook/IMAP. Every single company I've talked to wants this. Read, categorize, draft replies, file into folders. This is the universal use case.
- OneDrive/SharePoint integration — File access beyond local directories. The Knowledge Base feature works great locally, but enterprise documents live in the cloud.
- Webhook tool — Let agents trigger external workflows via HTTP. One company already has Power Automate pipelines; they just need Pinchy to kick them off.
- Mistral provider support — Multiple companies in the DACH region want a European LLM provider for GDPR reasons. Mistral is the obvious choice.
The integration architecture is taking shape too. I built an integrations tab this week — settings UI where admins configure connections (Odoo, soon Outlook, OneDrive) and a per-agent permission matrix for which connections each agent can use. Same allow-list philosophy as the tool permissions, applied to external systems.
The Business Model Conversation
We talked pricing. I'm still figuring this out, honestly. The community edition is free and fully functional — permissions, audit trail, the whole security stack. Enterprise adds groups and access control for larger teams.
For now, the deal with pilot companies is: you get everything free for three months, you give me aggressive feedback, and we figure out a fair price after that. Custom integrations (connecting Pinchy to your specific CRM or ERP) are separate — either I build it on a project basis, or I provide the docs for your team to build it.
Long term, the integrations need to be a plugin system that anyone can extend. I can't build connectors for every system. But right now, building a few myself teaches me exactly what that plugin API needs to look like.
Day 38
v0.3.0 shipped. A recruitment company is deploying Pinchy next week. A corporate innovation team wants to explore agent-to-agent communication. The roadmap went from "everything is possible" to "email, files, webhooks, Mistral — in that order."
I'm in an exploration phase. Talking to everyone, staying open to where this goes, resisting the urge to build the everything-machine. The plan: keep the core compact, make everything extensible — plugins, API, channels — and let the use cases guide the priorities. Ask me again in two months; everything changes fast. But the direction is clear: Pinchy should be something people can make their own, not something I have to customize for each customer.