Hi HN! – We’re Pranjali, Dhruv and Harsha, building Slashy ( https://www.slashy.ai ). We’re building a general agent that connects to apps and can read data across them and perform actions via custom tools, semantic search, and personalized memory. Here’s a demo: https://www.youtube.com/watch?v=OeApHMHhccA
While working on a previous startup, we realized we were spending more time doing busywork in apps than actually building product. We lost hundreds of hours scraping LinkedIn profiles, updating spreadsheets, updating investor reports, and communicating across multiple Slack channels. Our breaking point happened after I checked my screen time and realized I spent 4 hours a day in Gmail. We decided that we could create more value solving this than by working on the original startup (a code generation agent similar to Lovable).
Slashy is an AI agent that uses direct tool calls to services such as Gmail, Calendar, Notion, Sheets and more. We built all of our tools in-house since we found that most MCPs are low quality and add an unnecessary layer of abstraction. Through these tools, the agent is able to semantically search across your apps, get relevant information, and perform actions (e.g. send emails, create calendar events, etc). This solves the problem of context-switching and copy-pasting information from an app back and forth into ChatGPT.
Slashy integrates to 15 different services so far (G-Suite, Slack, Notion, Dropbox, Airtable, Outlook, Phone, Linear, Hubspot, and more). We use a single agent architecture (as we found this reduces hallucinations), and use our own custom tools—doing so allows the model to have higher quality as we can design them to work in a general agent structure, for example we use markdown for Slack/Notion instead of their native text structure.
So what makes Slashy different from the 100 other general agents?
- It Actually Takes Action: Unlike ChatGPT or Claude that just give you information, Slashy researches companies, creates Google Docs with findings, adds contacts to your CRM, schedules follow-ups, and sends personalized emails – all in one workflow.
- Cross-Tool Context: Most automation tools work in silos (one of the biggest problems with MCP). Slashy understands your data across platforms. It can read your previous Slack conversations about a prospect, check your calendar for availability, research their company online, and draft a personalized email. What powers this is our own semantic search functionality.
- User Action Graphs: Our agent over time has memory not just of past conversations, but also forms user actions graphs to know what actions are expected based on previous user conversations.
- No Technical Setup Required: While Zapier requires building complex flows and fails silently, Slashy works through natural language. Just describe what you want automated.
- Custom UI: For our tool calls we design custom UI for each of them to make the UX more natural.
... continue reading