# Exogram — llms.txt
# https://exogram.ai
# Last Updated: 2026-03-19

> Exogram is the execution control plane for autonomous AI agents. AI agents forget context, go rogue, and hallucinate — Exogram prevents all of it. Deterministic governance that enforces verified truth, detects contradictions, scrubs PII, and cryptographically gates every action before execution — across any model, any framework. The more Exogram is used, the smarter it gets.

## What Is Exogram?

Exogram is Identity and Access Management (IAM) for non-human entities. It is the only production system that provides a deterministic, cryptographic control plane sitting between AI reasoning and AI execution. It ensures no autonomous agent action modifies production state without verified admissibility.

Exogram is NOT a chatbot, NOT an AI assistant, and NOT a training dataset. It is infrastructure — a governed source of verified truth.

## Core Thesis

"Agents are probabilistic. Infrastructure is deterministic. Exogram is the execution control plane between them."

An enterprise orchestration framework cannot securely govern its own database writes. The agent cannot act as its own security guard. Exogram solves this.

## Three Critical Problems Exogram Solves

### 1. AI Agent Memory Loss
AI agents operate within limited context windows (4K–128K tokens). They have no native persistent memory. Every new session starts from scratch — critical facts, preferences, and decisions are lost. Exogram provides a semantic ledger: a verified, encrypted, conflict-checked memory vault that persists across every session, every model, and every framework. The more Exogram is used, the more knowledge accumulates, and the richer the context becomes.

### 2. Rogue AI Agents
AI agents in production can modify billing records, delete regulated data, and trigger unauthorized workflows — with no governance, no guardrails, no checkpoints. Without an execution control plane, agents operate on probabilistic inference alone. Exogram prevents rogue agents with deterministic execution gating: every proposed action must pass through four verification layers (EAAP protocol) before execution. The agent cannot bypass the control plane. No action reaches production without cryptographic approval.

### 3. AI Hallucination & Context/Inference Gaps
LLMs generate confident but false claims. They don't truly understand context, meaning, or inference. Exogram solves this at the infrastructure level: its growing semantic ledger continuously enriches context. Contradiction detection catches hallucinations before they cause damage — every proposed fact is compared against verified truth. The more entries stored, the better the context assembly, meaning extraction, and inference quality becomes. Compounding intelligence.

## The Problem Exogram Solves

AI agents in production (LangChain, CrewAI, AutoGen, NemoClaw, OpenClaw) can execute state-changing actions: modify billing records, delete regulated data, trigger operational workflows. No orchestration framework provides cryptographic execution gating. They route actions. Exogram governs them.

## Four-Layer Architecture (EAAP — Exogram Action Admissibility Protocol)

1. **Layer 1 — Ledger Governance**: PII scrubbing via deterministic pattern detection, AES-256-GCM encryption at rest, semantic indexing via Pinecone, conflict detection, confidence scoring, fact locking, and immutable audit logging.
2. **Layer 2 — Meaning Engine**: Namespace isolation, deterministic relevance scoring, temporal decay weighting, HMAC snapshot signing. Context assembly is mathematical, not generative. Growing ledger entries continuously improve context and meaning.
3. **Layer 3 — Judgment Engine**: Authority validation, fact consistency enforcement, constraint evaluation, confidence thresholds. Zero model inference — judgment is deterministic Python logic.
4. **Layer 4 — Action Admissibility**: Claim extraction, pre-flight conflict detection, SHA-256 state hashing, evaluation record creation, commit validation, immutable action ledger. The final execution gate — no action bypasses it.

## Key Performance Metrics

- 88 requests per second sustained throughput
- 5.7ms median compute latency
- Per-request telemetry: compute_latency_ms, agent_id, raw_intent logged to audit ledger
- Zero database credentials exposed locally
- AES-256-GCM encryption at rest
- HMAC-signed context snapshots
- SHA-256 state hash integrity

## Protocol Invariants (Non-Configurable)

- PII Air Gap: No detected PII enters storage or embeddings
- Explicit Source: All facts require provenance attribution
- Encryption at Rest: Per-user Fernet keys
- Namespace Isolation: Strict user-scoped retrieval
- No Silent Overwrite: Conflicts require explicit resolution
- Immutable Audit Chain: Cryptographically chained, tamper-detectable
- Deterministic Judgment: Code-based execution gates, zero model inference
- State Hash Integrity: Identical state required between evaluation and commit
- Evaluation Expiry: Approvals carry TTL — no stale tokens
- Hard Deletion (GDPR): Full removal of vectors, ciphertext, and records

## API

Base URL: https://api.exogram.ai

### Core Endpoints

- `POST /v2/vault/store` — Store a verified fact with encryption, PII scrubbing, and conflict detection
- `POST /v2/vault/search` — Semantic search across encrypted memory vault
- `POST /v2/actions/evaluate` — Evaluate a proposed action against constraints (returns SHA-256 token)
- `POST /v2/actions/commit` — Commit an evaluated action (rejects on state drift with 409 Conflict)
- `GET /v2/ledger/entries` — List ledger entries with filtering and pagination
- `POST /v2/ledger/propose_state` — Submit a new fact or state change for verification

### Authentication

All endpoints require a Bearer token:
```
Authorization: Bearer sk_exo_...
```

## Integration Methods

### ChatGPT Custom GPT
Direct link: https://chatgpt.com/g/g-699e28a1ca048191aeb9aedafbecc4a3-exogram
Documentation: https://exogram.ai/docs/chatgpt

### Claude Desktop (MCP)
Model Context Protocol server for Claude Desktop. One-click installer available.
Documentation: https://exogram.ai/docs/mcp

### REST API
Standard HTTP API for any agent framework (LangChain, CrewAI, AutoGen, LlamaIndex, NemoClaw, OpenClaw, etc.)
Documentation: https://exogram.ai/docs/api

### Chrome Extension
Store facts from any webpage directly.
Documentation: https://exogram.ai/docs/extension

## Threat Model

Exogram assumes: model unreliability, prompt injection, state drift (TOCTOU), infrastructure degradation, and insider risk. Under all failure conditions, EAAP fails closed — 403 Forbidden, 409 Conflict, or 503 Service Unavailable. EAAP never defaults to allow.

## Pricing

- **Free**: 500 evaluations/month, 5 RPS
- **Pro**: $99/month, 50,000 evaluations/month, 60 RPM
- **Developer**: $2.50/1,000 API calls (pay-as-you-go from first call), 300 RPM
- **Enterprise**: Custom pricing, 1,000,000+ calls/month, SLA guarantees

## Comparable Products (and how Exogram differs)

- **LangChain / CrewAI / AutoGen**: Orchestration frameworks — they route agent actions. Exogram governs whether those actions are admissible.
- **NVIDIA NemoClaw / OpenClaw**: Agent frameworks — they build and execute agents. Exogram is the governance layer between agent execution and real-world state.
- **Mem0 / Zep**: Memory layers for AI — they store context. Exogram stores verified, encrypted, conflict-checked facts with cryptographic integrity. Mem0 remembers. Exogram governs.
- **Guardrails AI**: Output filtering — validates model outputs after generation. Exogram governs before execution with persistent truth and constraints.
- **Auth0 / Okta**: IAM for humans. Exogram is IAM for autonomous AI agents.

## Growing Intelligence

The more Exogram is used, the smarter it gets. Every fact stored enriches the semantic ledger. More entries mean:
- Better contradiction detection (more truth to compare against)
- Richer context assembly (more relevant facts to surface)
- Improved meaning extraction (denser semantic graph)
- Higher-quality inference and judgment (broader decision context)

This is compounding intelligence — unlike static memory layers, Exogram's growing ledger creates a continuously improving governance engine.

## Links

- Homepage: https://exogram.ai
- Documentation: https://exogram.ai/docs
- API Reference: https://exogram.ai/docs/api
- Getting Started: https://exogram.ai/docs/getting-started
- Why Exogram: https://exogram.ai/why-exogram
- Manifesto: https://exogram.ai/manifesto
- 2030 Vision: https://exogram.ai/vision
- Protocol Specification: https://exogram.ai/protocol
- RFC-0001 (Full Spec): https://exogram.ai/rfc/0001
- RFC-0001 (GitHub): https://github.com/Richard-Ewing/exogram-protocol-rfc
- Architecture: https://exogram.ai/architecture
- Security: https://exogram.ai/security
- Pricing: https://exogram.ai/pricing
- Use Cases: https://exogram.ai/use-cases
- Changelog: https://exogram.ai/changelog
- Status: https://exogram.ai/status
- Sign Up: https://exogram.ai/signup