Launching today
Cleo
The AI PM that runs your team
230 followers
The AI PM that runs your team
230 followers
Cleo is the AI product manager for founders and lean teams. It lives in Telegram and Slack - earns your tone, knows your team, and runs the PM work (standups, follow-ups, decisions) while you ship the product. What's different: every fact Cleo learns is transparent - you see the source, the confidence, and can confirm or correct it. No black-box memory. Five trust levels, from observer to operator. Free in Telegram. 1 min setup
Cleo
@louislecat The endgame is bigger than a PM tool.
We're building toward agents that operate as employees - not assistants. They sit in your meetings. Execute across every integration your team uses (50 native, 500+ through Composio). Run entire functions of the company under your judgment.
Cleo today is the operational layer. The architecture (memory + provenance + learned rules + integrations) generalizes - same primitives that draft a client follow-up today will run the entire client-success function tomorrow.
An AI workforce that operates under human judgment, with auditable memory and trust gates. Not "AI that helps you." AI that does the work.
Kipps AI
Big Congrats on launch! Which AI models are you using? And how is company security handled?
Cleo
@nishit_chittora Thanks!
Models: 7 in production, routed by task - 3 from Anthropic (Claude, primary for agent reasoning and drafting in user voice), 2 from Google (Gemini, for long-context retrieval), 2 from OpenAI (GPT, for structured outputs). Routing via OpenRouter + LiteLLM, so we swap models per workload without changing app code.
Security: No cross-customer training (your data never trains shared models). Encryption at every message. OAuth-only integrations — we don't store passwords. Architecture made with SOC 2 in mind; certification in progress
Rizzle AI
What’s the biggest bottleneck you faced while building this?
Cleo
@nithin_raju1 The hardest problem wasn't model quality or integration count - it was keeping context coherent across tools without exploding cost or latency.
Really proud to have worked on the design here. The hardest thing wasn't making it pretty, it was making "you can see everything Cleo knows" not feel like homework. Source, confidence, a fix button on every fact… that's a lot to surface without burying people. Took a few rounds to get the chat and the memory side to live together cleanly. Happy with where it landed
Cleo
@islam_yuldashev The "not feel like homework" framing is the part I keep coming back to. You got it to feel like a Notion page that happens to know things, not a compliance dashboard. That distinction is the whole product working.
I've known Rahim since school. Watching him ship has been wild - they killed their own launch in May because telemetry wasn't right, rebuilt in in 5 days, and kept every paying customer. That kind of operating velocity is rare. Pulling for them
Cleo
@dedyfo Thanks for the kind words, Rovshen. The May 4 kill was the hardest call we've made so far - but the telemetry was honest about what was working and what wasn't.
What stands out about Cleo is that it doesn't just track tasks — it actually runs sprints autonomously. Curious how it handles prioritization when multiple stakeholders have conflicting requests. Does it have a built-in negotiation framework or just default to whoever shouts loudest?
Cleo
@lakshminath_dondeti The stack underneath is doing a lot - Supermemory abstraction, vectors for semantic recall, graph for entity relationships. The interesting part is that you never see it. You add a fact in chat, you ask a question across three projects, you get the right answer with source. That kind of "complexity stays under the hood" engineering is the real moat. Most AI products leak their architecture into the UX. This one doesn't.
Cleo
@lakshminath_dondeti Appreciate it. Curious - what made you ask about memory specifically? Building something adjacent, or evaluating Cleo for your own team?
Cleo
@lakshminath_dondeti Ha, that's the best phase to be in.