We built an AI workspace because our tools kept forgetting who we were

We all experience the same issue, every time we opened a new AI chat we have to re-explain at least one concept. The context that should have been obvious but unfortunately there isnt enough compute to take on continuous context. And then we'd close the tab, switch back to our task manager, our docs, our calendar, Apollo, posthog, stripe and so many other tools that maintain our competitive advantage, none of which your agents know anything about.

We were the glue between our own tools. Copying, pasting, summarizing, updating manually just to keep everything in sync. Across chats, kanbans, docs and so much more.

So we built Agently. Here's how it works:

• The Brain is the foundation. It's not merely a knowledge base, it's a context engine. It ingests your uploaded files, SOPs, brand guidelines, data from live integrations, and external AI apps like ChatGPT and Claude, then automatically refines, categorizes, and indexes everything. Every agent draws from it in real time, so responses are grounded in your actual business, your tone, your processes, your data not a generic output the internet has seen a million times.

• Spaces are Kanban boards where agents work alongside you like real teammates. Assign a task to an agent the same way you'd assign it to a person. Schedule tasks over time thresholds, or even allow agents to draft whole task plans to schedule, plan, assign and execute. It picks it up, executes it, and moves the card. Don't be surprised when an agent assigned you a task rather than vice versa.

• Pages is a full document operating system: PDFs, HTML files, sheets, presentations, rich documents. It ships with a Notion-style template library, and agents can generate new templates on demand. It's where your team's knowledge gets created, not just stored. You can choose what to add to the brain.

• Integrations: 150+ MCP-powered connectors covering the tools founders actually use. Two-way. Agents don't just read your stack, they write back to it. Not "connect your calendar." Actually act on it.

• Chat works like Slack, humans and agents in the same threads, context is summarized and categorized into brain

• Calendar pulls from everything you connect into a single unified view.

• Your AI Cofounders, agents with continuous context whether long term context stored in the brain or short term trailing checkpoint memory ensuring continuous memory with wide token input.

And we're building toward Jarvis: a lead agent that orchestrates the entire workforce on your behalf, delegates across functions, and keeps everything moving without you managing it manually.

We're early and we want honest, direct feedback from founders and small teams who've lived this problem.