Navigating the agentic future.
AgentsDB is a new project dedicated to covering the fast-evolving landscape of AI agents, product launches, ecosystem players, and the trends shaping tomorrow’s intelligent systems. Our mission is to make sense of this transformation and guide you through it with clarity and depth.
What's happening today?
Comet’s background assistant makes the browser an agent
Perplexity Comet’s background assistant is now public, showing how the browser can host real agents that plan, act, and ask for consent before they click. Here is what changes, why it matters, and how to try it today.
Neo4j bets $100M to make graph memory the agent norm
On October 2, 2025, Neo4j unveiled Aura Agent in early access and a native MCP server, backed by a $100M investment. The goal is to move enterprise agents from brittle vector recall to persistent, explainable graph memory.
Instant Checkout Goes Live: ACP Turns Chat Into Commerce
On September 29 to 30, 2025, OpenAI and Stripe introduced Instant Checkout and the Agentic Commerce Protocol. Learn how ACP standardizes agent to merchant transactions, what changes for Etsy and Shopify sellers, and how to adopt it in 90 days.
When Web Pages Become Workers: The Agentic Browser Era
AI is moving from chat boxes into the browser itself. Opera’s Neon can act on pages, Perplexity made its Comet browser free, and Google is wiring Gemini into Chrome. The page is turning into a capable colleague that gets real work done.
ShinkaEvolve and the rise of test-time evolutionary compute
Sakana AI's ShinkaEvolve flips the script on scaling by evolving code and training signals at test time. Novelty filters, smart parent sampling, and model bandits deliver state-of-the-art results with far fewer trials.
Boomi brings MCP to Agentstudio, the USB-C for agents
Boomi’s September 2025 update brings Model Context Protocol to Agentstudio, turning existing connectors into safe, schema-based tools that agents can discover, trust, and govern. See what it unlocks now and how to adopt it with confidence.
Microsoft Security Store puts autonomous agent teams in charge
Microsoft’s Security Store pairs a no code builder with a curated marketplace so SOCs can deploy coordinated AI agents that investigate, enforce, and learn under guardrails. Here is how to evaluate, adopt, and govern the shift.
Memory Is the New Moat: Context Becomes Capital in AI
As models converge, memory diverges. This essay shows why portable, governed context becomes the real moat for agents, and how to design wallets, grants, receipts, and ledgers that compound product value.
OutSystems Agent Workbench GA makes enterprise AI agents real
OutSystems has taken Agent Workbench to general availability, formalizing an agent workbench category: a low code control plane to design, govern, and deploy multi agent systems across live enterprise data and workflows with real controls.
Discovery Becomes Runtime: AI Drugs Move From Code to Clinic
AI drug discovery is shifting from slow handoffs to a live runtime that compiles hypotheses, runs experiments on robots, and streams evidence to regulators. Funding, human data, and agentic tools point to a stack that is already deploying.
Ray3 lands: Luma's reasoning video model goes production-grade
Luma's Ray3 is now inside Adobe Firefly, bringing native HDR, ACES EXR sequences, and a fast Draft Mode that preserves composition. With Creative Cloud handoff and agency partners, AI video moves from demo to deliverable.
Vertex AI Agent Engine’s September leap to real runtime
September 2025 turns Vertex AI Agent Engine into a production ready runtime with sandboxed code execution, agent to agent collaboration, durable memory, bidirectional streaming, and tightened enterprise controls.
Likeness Liquidity: Renting Your Face After Sora’s Surge
Sora's breakout week turned likeness into a tradable asset. This guide shows how consent, pricing, and watermark rails turn presence into a market, and what studios, unions, creators, and platforms must do next.
Tinker brings distributed fine-tuning to open models
Thinking Machines Lab introduced Tinker on October 1 and 2, 2025, a developer-first service that turns distributed fine-tuning of open-weight LLMs into a turnkey step. The promise is sovereign AI and scale without owning a GPU farm.
Claude Sonnet 4.5 pushes agents from demos to dependable work
Flashy demos are over. Claude Sonnet 4.5 pairs accurate computer use with long unattended runs and shipping-grade scaffolding, so teams can move from pilots to production agents that meet real service levels.
Opera Neon and the Dawn of the Agentic Browser
Opera Neon makes the browser an on-device agent that reads the page, fills forms, and completes jobs you can audit. We compare Neon, Comet, and Dia, then outline what enterprises must ship in the next year.
Idle Agents Are Here: Systems That Keep Working Offscreen
A late September surge of agent updates marked a shift from chat-first tools to time-native systems that keep working between check-ins. Learn how open-loop design reshapes UX, pricing, and trust.
GitLab Duo Agent Platform hits beta, DevSecOps orchestrated
GitLab turned agentic development into production reality. Duo Agent Platform enters public beta with IDE and web chat, an orchestrated Software Development Flow, MCP support, and integrations for JetBrains and Visual Studio.