Install the Fulcra skill to give your agent real world context about your life.
For the first time, experience a personalized agent who knows your ground truth .
Hermes remembers what it did, evolves its skills, and builds procedural recall from every session. But what it cannot do is reach the data that lives outside the conversation like your calendar today, last night's sleep, your HRV, where you are, what you're taking. Fulcra is that real world knowledge layer. Install the Fulcra skill, configure once, and your Hermes agent can query your real-world state on demand without ever logging into any of your accounts, with access that's scopeable and revocable.
Google Calendar, Apple Calendar, Apple Health, Garmin, Oura, Whoop, Dexcom, nutrition apps, and your location throughout the day. Everything your wearables and apps already create, unified into a single source of truth your Hermes agent can query. Install the Fulcra skill, connect your sources once, and your agent has access without account logins or manual context-pasting.
Hermes's MEMORY.md holds what your agent has learned: your preferences, your projects, your environment. Fulcra holds what's true about your life right now including your sleep stats, your HRV, your calendar load, and your current medications. These change daily and are too large and dynamic for MEMORY.md. Fulcra gives your agent a place to query them on demand.
Hermes's MEMORY.md is capped at ~2,200 characters. Real-world state — biometrics, calendar, location — doesn't belong there. It's too large, it changes every day, and it ages out the moment it's written. Query it live from Fulcra instead, scoped to what the current task needs, and keep your memory focused on what it's designed for: durable facts about your environment and preferences.
Hermes writes and edits its own skills through use. Fulcra is a ground truth reference that is immune to drift, making it reliable data you can trust, no matter how much your agent evolves.
When Fulcra integrates with Hermes, your agent goes from helpful chatbot to a true utility that actually knows you. Here is what that looks like in practice.
When you wire Fulcra into your Hermes build, each subagent can query the same source at call time and pull only what's relevant to its task. A research subagent gets your current focus window and timezone to scope its output. A scheduling subagent gets your calendar load and sleep quality to protect your afternoon. Both query the same Fulcra snapshot. Neither gets the other's data.This maps directly to how Hermes handles delegation: subagents are isolated by design, receiving only the goal and context the parent passes them. Fulcra gives you a clean way to scope real-world data to each subagent rather than bloating the parent context with everything.
"Subagent A is researching Q2 market shifts — it knows your focus window is clear until 2pm. Subagent B is protecting your afternoon — it knows your sleep quality was low and your next hard block starts at 3. Same Fulcra data source, scoped per task."
→ See how Hermes handles subagent delegation: Delegation & Parallel Work
The Daily Briefing Bot is the canonical first Hermes build. By default it pulls news, weather, and RSS: what’s going on with the world. With Fulcra, it pulls the data about what’s going on with you. Use for biometric Health detection and more.
Try it now by pasting this into Hermes: "Use Fulcra to pull my sleep stats, HRV, today's calendar, and any active health experiments. Then give me my morning briefing: what I should front-load, what I should protect, and what I need to know before my first meeting."
"You slept 7.2h (with less deep sleep than the night before). HRV is up 8ms — good trend. You have 4 meetings today, your glucose experiment is on day 3, and your run is overdue by 2 days. Front-load the hard cognitive work before noon, protect the 2pm block, hold the line on the experiment. I'll have your evening numbers ready at sundown."
→ Follow the full Hermes Daily Briefing Bot tutorial: Tutorial: Daily Briefing Bot
Once Fulcra is wired in, your Hermes agent can form hypotheses and find patterns across your biometrics, behaviors, and inputs over months and across data streams no single wearable can correlate alone.
Try it now by pasting this into Hermes: "Use Fulcra to pull my HRV, sleep, calendar, and any supplement or medication logs for the past 6 weeks. Look for patterns I haven't noticed. What's hurting my recovery? What's helping?
"Your HRV dropped 5ms, but it's Sunday, the day after your injection, and you mentioned drinks Saturday. Same pattern 3 of the last 4 weeks. Want to test it? Skip Saturday drinks this weekend and I'll watch whether Sunday HRV holds."
Location, HRV, glucose, sleep, calendar, plus your own annotations: medications, supplements, devices, mood. Your Hermes agent finds the patterns across all of it.
Try it by pasting this into Hermes: "Use Fulcra to pull my sleep, HRV, glucose, and any device or supplement logs for the past month. Find correlations between what I'm logging and how I'm recovering."
"On days you log your Elemind headband and less than 1 hour of Netflix, deep sleep averages 28% vs 12% without. Coffee before 10am: afternoon HRV 15% higher than coffee after noon."
People move between tools. You might run Hermes for a while, try something else, or use different agents for different tasks. The problem is your personal context like, your health data, calendar patterns, and preferences, live in whichever agent you were last talking to, or nowhere at all.
Fulcra sits outside any single agent. Your data is there when you open Hermes, and it would be there in any other tool you connect to Fulcra.
"Think of it like the difference between saving your work in one app vs. saving it to disk. Fulcra is the disk."
To install the Fulcra skill, run this at your terminal:
hermes skills install fulcradynamics/agent-skills/skills/fulcra-onboarding
/fulcra-onboarding help me set up a new personal dashboard
Fulcra is an infrastructure layer that improves agent performance by giving your agent real world knowledge about you. We are another, necessary, type of agent knowledge like agent memory.
vs. Hermes Native Memory (MEMORY.md, USER.md, session search)
Hermes memory stores what your agent has learned through conversation, like your preferences, your project conventions, past sessions. Fulcra stores all the conditions of your life right now: your sleep from last night, today's calendar, your current HRV, what you're taking. These are too dynamic and too large for MEMORY.md. They belong in Fulcra, where they are queryable on demand.
vs. Memory Providers (Honcho, OpenViking, Mem0, Hindsight, Holographic, RetainDB, ByteRover, Supermemory)
Memory providers extend what Hermes can remember about your conversations with semantic search, knowledge graphs, cross-session recall. Fulcra is complementary: it's not a memory store for conversations, it's a data source for your real-world state.
vs. Apple Health / Whoop / Oura
Wearables give you one data stream each, in their own app, in their own format. Fulcra unifies all of them into a single source your Hermes agent can query. One auth, one access point, across everything you already track.
Fulcra was designed by people who get privacy and know the importance of an infrastructure solution that can be the secure private datastore for the rest of your life. Here data is yours, under your control, and only shared with the people and tools you choose to share it with.
