Blog
Articles, insights, and deep dives from the Arcanada team.
Adsessor: Introducing an In-Call AI Assistant for Zoom and Google Meet
We are building Adsessor — an AI assistant that listens, transcribes, translates, and eventually speaks during your Zoom and Google Meet calls. Phase 0 design validation is in progress; here is the full picture: the problem, the architecture, the roadmap, and why we are building it in public.
7.7× Cheaper and 4× Faster: Splitting Inference Across Claude, Kimi, and DeepSeek
I build Datarim, a framework for AI agent pipelines, and two days ago I added DeepSeek to the stack as the third model alongside Claude and Kimi K2.6. One thinks, one writes, one reads and summarizes.
The Scanner Manifesto: Big System vs. Discipline
Scanners get told to focus their entire lives. Maybe the problem isn't that they switch — it's that they don't have a system big enough to make switching useful. One person, five servers, twenty repos, ~30 domains, 266 closed tasks in three weeks. AI agents change the economics of being a scanner.
From Telecom to AI Agents: a Six-Level Autonomy Scale That Already Exists
I read Kirill Simakov on autonomy levels for agent systems. The framing was useful but the post had no references, so I went looking for the source. The trail led to TM Forum’s Autonomous Networks framework, with parallel material from Ericsson, McKinsey, and Gartner. The same scale that telecom operators have used for years maps onto AI agents almost cleanly. I applied it to the Arcanada ecosystem and put the ratings on the public site.
AI Agent Model Connector Benchmark: CLI Subscriptions vs API Providers
I benchmarked 20 AI models across 5 connectors. CLI tools score 100% accuracy but are forced to run sequentially — 20x slower than parallel API calls. Here are the numbers, the architectural tradeoffs, and my decision for the Arcanada ecosystem.
Scrutator: Building the Knowledge Engine Behind Arcanada
We built Scrutator — an open-source search, retrieval, and meaning engine. Dense + sparse + ColBERT embeddings, hybrid search, and a Dreaming module that reorganizes knowledge while you sleep. All five layers are live in production.
How I Picked an Embedding Model for Arcanada
Benchmarked 7 multilingual embedding models on CPU, no GPU. bge-m3 won with 0.887 cross-lingual similarity — 45% higher than the nearest competitor. Here's the full breakdown.
Datarim: Do Agents Dream of Electric Sheep?
AI agents can write code. They can't run projects. Datarim is an open-source framework with 15 agents, 18 skills, and a 9-stage pipeline that gives AI agents project execution structure. 1,000+ tasks closed in one year by one person.
Claude Code Memory Optimization: Auto Memory, Auto Dream, and the Bugs Documentation Doesn't Mention
Claude Code has a problem you'll only discover after 20-30 sessions: its memory degrades. Here's how Auto Dream is supposed to fix it — and why the /dream command doesn't actually work.
Introducing Arcanada: A Comprehensive AI Agent Ecosystem
We are thrilled to announce the official launch of Arcanada — a comprehensive ecosystem for building, orchestrating, and scaling AI agents and swarms.