{"slug": "i-stopped-re-explaining-my-project-to-every-ai-model-here-s-what-changed", "title": "I stopped re-explaining my project to every AI model. Here's what changed.", "summary": "Kanchan Satyal, co-founder of software studio 123sudo, built 9xchat, a workspace that keeps persistent memory across cloud and local AI models so context and decisions carry between tools and teammates. He reports that saved decisions, winning post formats, and past conversations persist across model switches, though he notes memory still requires manual updates when facts change. The product offers automatic model routing via a GPT-based selector, manual model choice, bring-your-own API key, and shared conversations and agents.", "body_md": "I'm Kanchan satyal , co-founder of 123sudo. We're a small studio that makes everyday software: 9xchat, 9xbuddy, 9xconvert, and 9xlyrics.\n\nRunning a portfolio like that means my days don't stay in one lane. In one afternoon I might write a Threads post, check a product decision, draft a YouTube description, and look at a weekly analytics recap.\n\nFor a long time, every one of those tasks began the same way: **explaining who we are again.**\n\nIf you use more than one AI tool, you probably know it:\n\nNone of the models were bad at their jobs. The workflow around them was broken. I was the only thing connecting them, and I was doing it by copying and pasting.\n\nRe-explaining your context isn't one paragraph. For us it includes:\n\nContext like this changes all the time. A product gets dropped. A post format starts working. A rule gets updated. Keeping it current across different tools by hand never worked for long.\n\nWe run 123sudo's growth work inside 9xchat now: strategy, Threads replies, captions, research, and planning. Work is sorted into projects for 123sudo, 9xchat, and marketing.\n\nHere's what changed in practice:\n\n**1. It remembers our decisions.**\n\nWhen we removed a product from our portfolio, that decision was saved once. I don't have to bring it up in every new chat.\n\n**2. Our wins get written down.**\n\nWhen one Threads format clearly beat the others, the pattern got saved: an industry-wide hook, a quick list of examples, what it costs the user, what we do differently, a strong close. The next draft builds on it instead of starting from zero.\n\n**3. I can find old work again.**\n\nI once needed a caption we'd written weeks earlier for a 9xbuddy post. I didn't dig through tabs. I asked, and it found the original chat and the exact text.\n\n**4. I can let Auto choose, or choose for myself.**\n\nFor routine work, I use Auto. A GPT model selects a model for the task and routes it there, so I don't have to make that decision every time. When I want a specific model, I choose it manually. 9xchat also lets users bring their own API key.\n\nBecause memory carries across models, switching doesn't mean starting from a blank page.\n\n**5. My teammate can pick up where I left off.**\n\nWhen I need to hand work over, I share the 9xchat conversation with a teammate. They continue in that same chat, with the discussion already there. I can share agents too, so they use the same setup instead of rebuilding it.\n\nIt feels less like passing around a document and more like handing off an active workspace.\n\nMemory isn't magic, and I don't want to oversell it.\n\nWe changed our welcome offer from $1 worth of AI tokens to 1 million AI tokens. Yet the assistant kept mentioning the $1 offer, even after I corrected it.\n\nApparently, I'd built a workspace that remembered our brand a little too well.\n\nPersistent memory isn't the same as automatically keeping every fact current. When something changes, I update the memory and check that drafts reflect it. Memory reduces repetition. It doesn't remove the need for judgment.\n\nBetter models matter. But for daily work, the bigger difference came from a workspace that remembers what I already told it.\n\nA smarter model that forgets you every time still makes you do the remembering. I'd rather spend that time shipping.\n\nThat's why we built 9xchat: one workspace, cloud and local models, persistent memory across them, and transparent credits.\n\n[Try 9xchat. Start instantly with 1 million AI welcome tokens, no API key required.](https://123sudo.com/desktop/9xchat?utm_source=devto&utm_medium=blog&utm_campaign=reexplaining)", "url": "https://wpnews.pro/news/i-stopped-re-explaining-my-project-to-every-ai-model-here-s-what-changed", "canonical_source": "https://dev.to/katsudo/i-stopped-re-explaining-my-project-to-every-ai-model-heres-what-changed-4hol", "published_at": "2026-10-06 03:09:23+00:00", "updated_at": "2026-10-06 03:17:43.773852+00:00", "lang": "en", "topics": ["ai-products", "ai-tools", "large-language-models", "ai-agents"], "entities": ["Kanchan Satyal", "123sudo", "9xchat", "9xbuddy", "9xconvert", "9xlyrics", "Threads", "YouTube"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/i-stopped-re-explaining-my-project-to-every-ai-model-here-s-what-changed", "markdown": "https://wpnews.pro/news/i-stopped-re-explaining-my-project-to-every-ai-model-here-s-what-changed.md", "text": "https://wpnews.pro/news/i-stopped-re-explaining-my-project-to-every-ai-model-here-s-what-changed.txt", "jsonld": "https://wpnews.pro/news/i-stopped-re-explaining-my-project-to-every-ai-model-here-s-what-changed.jsonld"}}