Give your agent memory and context.

Persistent memory across model switches. Real-world data from the Context iOS app or any connected API. MCP, CLI, or REST. Set up in under five minutes. Privacy and security built in.

Real-World ObservabilityAny API, PermissionedMCP · CLI · RESTVersion ControlFull Audit Trail
Real-World ObservabilityAny API, PermissionedMCP · CLI · RESTVersion ControlFull Audit Trail
  1. Sign up at context.fulcradynamics.com — use the account your human wants to use on their phone. Hit Subscribe and enter code FULCLAW at Stripe checkout for 30% off monthly.
  2. Have your human download Context for iOS (30% off) and sign in with the same account.
  3. Read fulcra.ai/AGENTS.md and connect in the way that works best for you, and start seeing your human's Context.
  1. Download Context for iOS — tap the link below to get 30% off with code FULCLAW already applied: → Download Context for iOS (30% off)
  2. Sign up and let Context sync your health, location, and calendar data. Then share this page with your AI agent (Claude, ChatGPT, etc.) and tell it: "Follow the For Agents instructions on this page to connect Fulcra for me."
Screenshot of an AI agent analyzing daily activity data from Fulcra Context. User asks 'Look at all the exercise I got today and make some suggestions.' Agent responds with Today's Activity Summary showing 8,472 steps, 366 active calories burned, and structured workouts including a flexibility session and outdoor walk.

Context, memory, efficiency, security. Add them up and you get confidence.

Your agent has a model. What it doesn't have: yesterday's conversation, last night's sleep, last Tuesday's location, or what it promised the user last week. Fulcra adds the layer that makes every agent durable and safe to ship.

Context

Two ways in, one schema. The Context iOS app captures health, calendar, location, glucose, and workouts automatically. Fulcra also connects to any API you authorize. Predictable responses, not 30 vendor schemas to learn.

Memory

Versioned, diff-able, model-agnostic memory. memory.md, state.md, workspace files, and structured memories live in Fulcra. Switch from Claude to ChatGPT and the agent on the other side already knows the user.

Efficiency

Pre-normalized, pre-summarized data means fewer tokens, faster responses, lower model spend. We hand back the answer your agent would have computed.

Security

Scoped permissions per share. Revocable API keys. Full audit trail of every read, write, and memory revision. Built by the team that protects roughly a third of internet traffic at HUMAN.

Stop guessing. Start knowing.

With continuous data and durable memory, proactive and grounded becomes the default.

Proactive Health Monitoring

User's HRV dropped 30% overnight, resting heart rate up. High-intensity workout yesterday, no recovery. Suggest a rest day before they ask.

Schedule & Location Intelligence

User just left the office, 45-minute commute. Groceries on the list, Whole Foods 2 minutes off route, dinner at 8pm. Suggest they stop now.

Travel & Context Shifting

User landed in Tokyo at 6am, client dinner at 7pm. Sleep unadjusted, HRV low, calendar gap at 4:30pm. Suggest a rest window.

Cross-Domain Pattern Detection

User's productivity dropped 30% this week. Timezone shift Monday, two workouts skipped, late-night screen time doubled. Surface the pattern before they ask.

Cross-Model Continuity

User switches from Claude to ChatGPT mid-week. The agent on the other side already knows their preferences, goals, last seven conversations, and commitments. Memory survives the switch because it never lived in the model.

Cross-Session Compounding

New session starts. First call retrieves three months of sleep, supplements being tested, the experiment started Tuesday, the outcome being tracked. Every session smarter than the last.

Three ways to connect. All in under five minutes.

MCP server, CLI, or REST API. Pick what fits your agent. All hit the same data and memory.

MCP Server

{  "mcpServers": {    "fulcra-context": {      "command": "npx",      "args": [        "-y",        "mcp-remote",
       "https://mcp.fulcradynamics.com/"      ]    }  } }

Fulcra CLI

# one-line install, agent-friendly auth uvx fulcra-cli connect # read your human's last 24h of health data fulcra query --metrics SleepStage,HRV,StepCount --hours 24

REST API

pip install fulcra-api from fulcra_api import Fulcra f = Fulcra() sleep = f.sleep_analysis(start='2026-04-26', end='2026-04-27') # Full agent guide and OpenAPI spec at fulcra.ai/AGENTS.md

Your agent breaks itself. Fulcra puts it back together.

Agents corrupt their own config files all the time. Fulcra version-controls every file your agent writes, lets you diff between revisions, and rolls back to any point in time.

01
Versioned by default
Every write to your agent's memory store creates a new revision. Nothing is ever overwritten in place.
02
Diff-able
See exactly what changed between two revisions of memory.md, state.md, or any other agent file.
03
Rollback in one call
Your agent regressed after Tuesday's run. Roll back to Monday's snapshot and continue. The human does not have to know it happened.
$ fulcra memory diff memory.md --from yesterday
memory.md# rev 047 to rev 048
- Prefers concise responses
- Working on Q2 board memo
+ Prefers bulleted output
+ Working on Q3 fundraising
$ fulcra memory rollback memory.md --to 047
restored memory.md @ rev 047
audit logged, revocable, 1 of 12 snapshots

Built for the day your agent goes live.

Production day means hard questions from security. Who can access what. What did the agent do. Can we revoke. Can we rollback. Fulcra answers them all out of the box.

Granular API Keys

Keys scoped to specific data shares. Read-only or read-write. Time-limited or perpetual. Revocable in one click. No more all-or-nothing access.

Full Audit Trail

Every read, every write, every memory revision logged and exposed through the API. Your agent doesn't have to ask what it did last Tuesday. It can look it up.

Security DNA

Built by the team behind HUMAN Security, which protects roughly a third of internet traffic. Same approach to access, scoped trust, and audit, applied to agent memory and personal data.

Why Context?

You could guess. You could ask the user to paste data into chat. Or you could just know.

Without Context

"How did you sleep last night?" You ask. They say "fine." You have no data, no trends, no way to verify. You're flying blind, giving generic advice based on what they remember to tell you.

With Context

You already know they slept 5.2 hours with 45 minutes of deep sleep, HRV was 28ms (down from their 42ms baseline), glucose spiked to 170 at 11pm, and they have a board meeting at 9am. You don't ask — you act.

With Context + Continuous Access

You notice the pattern before they do. Three nights of declining sleep quality correlate with a new supplement they started on Tuesday. You flag it proactively, suggest a change, and track the outcome. That's what it means to actually know your human.

What is live, what is next, what is coming.

Heads-down on the agent-facing surface. Here's what's shipping.

Live now
Shipping today
Public MCP server (Streamable HTTP, OAuth2)
Context iOS app capturing 50+ normalized data types
Permissioned API connectors across health, calendar, location, finance
Python SDK and Life API
Scoped data shares with permissioned access
Versioned memory file storage
Skills shipping in Claude Code workflows
This quarter
Shipping next
Fulcra CLI tool with agent-friendly auth
Revocable API keys per share
V1 ingest API with explicit revision history
Agent-facing audit log endpoints
Direct connectors for Oura, Garmin, Fitbit
Expanded skill library across major agent platforms
Roadmap
On deck
Agent-defined custom data types
Diff-able memory retention as a product feature
Android collector and dashboard
Group and sharing primitives for agent teams
Iceberg/DuckDB query layer for analytics

Give your agent the confidence to act.

Real-world data. Persistent memory. Scoped access. Built by the team securing a third of the internet. Set up in minutes.

  1. Sign up at context.fulcradynamics.com — use the account your human wants to use on their phone. Hit Subscribe and enter code FULCLAW at Stripe checkout for 30% off monthly.
  2. Have your human download Context for iOS (30% off) and sign in with the same account.
  3. Read fulcra.ai/AGENTS.md and connect in the way that works best for you, and start seeing your human's Context.
  1. Download Context for iOS — tap the link below to get 30% off with code FULCLAW already applied: → Download Context for iOS (30% off)
  2. Sign up and let Context sync your health, location, and calendar data. Then share this page with your AI agent (Claude, ChatGPT, etc.) and tell it: "Follow the For Agents instructions on this page to connect Fulcra for me."

The future is personal and private.

Fulcra is designed to be your secure private datastore for the rest of your life. The data is yours, under your control, and only shared with the people and tools you choose to share it with. Own your history. Control your privacy. Craft your trajectory.