{"slug": "fulcra-wants-your-ai-agent-to-talk-to-anyone-else-s-no-matter-whose-platform-re", "title": "Fulcra Wants Your AI Agent To Talk To Anyone Else's, No Matter Whose Platform They're On", "summary": "Fulcra Dynamics, a Boston-based company building a user-owned context backend for AI agents, announced universal agent-to-agent communication that lets users of any agent harness — including Claude, ChatGPT, Hermes or OpenClaw — connect with another person's agent to coordinate shared tasks without either side switching platforms or creating a new account. CEO and co-founder Michael Tiffany framed the pitch against Instinct, whose agents talk only to other Instinct agents, saying \"locking in users provides a short term sugar rush that will ultimately crash when people realize they want and need more interoperability.\" The feature sits on top of Fulcra's context infrastructure layer, which normalizes a user's data onto a timeline and grants each agent access only to the context and actions its owner has explicitly authorized.", "body_md": "*\"Today's best agent may not be tomorrow's best agent. Users should be able to choose the agent that works best for a particular task and still have that agent access the context they own and communicate with other agents on their behalf,\" said Brandon Creighton, CTO of Fulcra Dynamics.*\n\nFulcra Dynamics, a Boston-based company that builds a user-owned context backend for AI agents, has announced universal agent to agent communication, a feature that lets anyone using any AI agent harness, including Claude, ChatGPT, Hermes or OpenClaw, connect with another person's agent to coordinate shared tasks. Neither side of the exchange has to switch platforms or set up a new account to take part, which is the detail Fulcra is betting matters most as the personal AI market gets crowded.\n\n## Why interoperability is the whole pitch\n\nMost people who have used agent to agent communication so far have done it through Instinct, where Instinct agents talk only to other Instinct agents. Fulcra's founders describe that as the wrong model for where the category is headed, since it only works if everyone standardizes on the same platform. That kind of lock-in narrows choice and raises trust and security questions for anyone who would rather not hand a new AI product access to their personal data just to make plans with a friend. It also puts the burden of adoption on both sides of any exchange at once: a friend, a colleague, or a business contact has to be talked into a specific product before two agents can even start coordinating, which is a much higher bar than simply asking whatever assistant someone already has open to reach out on their behalf.\n\n\"Instinct showed people that agent to agent communication can dramatically improve some aspects of our lives, but locking in users provides a short term sugar rush that will ultimately crash when people realize they want and need more interoperability,\" said Michael Tiffany, CEO and co-founder of Fulcra Dynamics. Fulcra's answer is to sit underneath whatever agent someone already uses rather than replace it, brokering the exchange and giving each agent access only to the context and actions its owner has explicitly authorized. That framing treats the agent itself as replaceable and the context layer underneath it as the durable asset.\n\n## What the product is actually doing underneath\n\nThe multiplayer feature sits on top of Fulcra's broader product: a context-focused infrastructure layer that pulls in a user's data, normalizes it, and puts it on a timeline. That gives agents a consistent way to query for what has changed in someone's life and act on it, whether the workflow is human to agent or agent to agent. It is that shared context, more than the messaging itself, that Fulcra argues is the hard part of making multiplayer AI actually useful instead of just another chat integration. A message passed between two agents is only as useful as the context each one can bring to it, so the timeline underneath is what lets an agent understand a request instead of just relaying it.\n\n[AI bots now make up half of traffic to the Fed's FRED data site](https://startupfortune.com/ai-bots-now-make-up-half-of-traffic-to-the-feds-fred-data-site/)\n\nFederal Reserve Governor Christopher Waller says AI agents now drive about half of all visits to FRED, the Fed's 850,000-series economic database, prompting a new MCP Connector built for machine readers instead of browsers. - [how to get economic data for AI models](https://startupfortune.com/ai-bots-now-make-up-half-of-traffic-to-the-feds-fred-data-site/) - [Federal Reserve FRED database AI bot traffic surge](https://startupfortune.com/ai-bots-now-make-up-half-of-traffic-to-the-feds-fred-data-site/)\n\nThe company has pointed to a few concrete examples of what that looks like in practice. In one, Fulcra's own marketing team took a bot that co-founder Ash Kalb had originally built to monitor Reddit conversations, connected it through Fulcra, and re-tooled it to be more effective for the company's marketing stack, an internal illustration of an agent built for one job getting reused for another once it can share context. In another, described as Kalb's favorite example, the same interoperability lets a person manage an entire fleet of agents working on a shared task rather than babysitting each one separately. Michael Tiffany's preferred example is a simpler one: scheduling a dinner between friends, or building a shared workspace for a new project, without either person switching tools to make it happen. Taken together, the three examples point in the same direction: an agent that was purpose-built for one narrow job can be redirected without being rebuilt, a single person can coordinate several agents instead of switching between them, and two separate people's agents can resolve a shared logistics problem without either person acting as the go-between.\n\n## Built by people who spent years protecting the rest of the internet\n\nFulcra says it was built over four years by a team of security-focused engineers who previously worked on VeraCode and Human, and whose work protects more than 30% of the internet from bot traffic. That background shows up in how the company frames the product's guardrails: users can see what any agent knows, trace how it learned it, and revoke access at any time, and agents coordinate through the shared layer without ever exchanging credentials or impersonating the person they represent. For a category where the core anxiety is handing a new piece of software broad access to personal data and history, that pedigree is part of the pitch as much as the interoperability itself: the people building the permission layer are the same people who spent years building systems meant to detect and stop unauthorized automated access.\n\nThe company is explicit that it wants competitors to Instinct to build on top of Fulcra rather than start separate walled networks. \"We want the people building alternatives to Instinct to build with us,\" Tiffany said. \"Make an agent people love. Let it work with the agents their friends love. I think that gives all of us a much bigger opportunity than trying to recruit everyone into separate networks.\"\n\nThat framing matters because, as CTO Brandon Creighton puts it, \"The AI agent market is changing too quickly for users to build their lives around one agent.\" Fulcra's bet is that the company that wins isn't the one with the best single agent, but the one that lets people's context and relationships survive a switch between agents entirely. Anyone who wants to see it firsthand can point an existing agent at fulcradynamics.com to get started, without needing to adopt a new one first.\n\n## Join the discussion\n\n[Open in the community →](https://startupfortune.com/community/)\n\nAlmost there. Sign in and your reply posts straight away.", "url": "https://wpnews.pro/news/fulcra-wants-your-ai-agent-to-talk-to-anyone-else-s-no-matter-whose-platform-re", "canonical_source": "https://startupfortune.com/fulcra-wants-your-ai-agent-to-talk-to-anyone-elses-no-matter-whose-platform-theyre-on/", "published_at": "2026-10-07 10:57:58+00:00", "updated_at": "2026-10-07 11:17:32.731015+00:00", "lang": "en", "topics": ["ai-agents", "agent-protocols", "artificial-intelligence", "ai-products"], "entities": ["Fulcra Dynamics", "Brandon Creighton", "Michael Tiffany", "Ash Kalb", "Instinct", "Claude", "ChatGPT", "OpenClaw"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/fulcra-wants-your-ai-agent-to-talk-to-anyone-else-s-no-matter-whose-platform-re", "markdown": "https://wpnews.pro/news/fulcra-wants-your-ai-agent-to-talk-to-anyone-else-s-no-matter-whose-platform-re.md", "text": "https://wpnews.pro/news/fulcra-wants-your-ai-agent-to-talk-to-anyone-else-s-no-matter-whose-platform-re.txt", "jsonld": "https://wpnews.pro/news/fulcra-wants-your-ai-agent-to-talk-to-anyone-else-s-no-matter-whose-platform-re.jsonld"}}