🧠 Memory Engine · Layer L501

Persistent, searchable memory for every AI agent and user session.

Store, compress and recall context across sessions — so your agents remember users, preferences and past decisions. Redis-backed, cryptographically sealed, 78p back on every £1.

Start building free View docs Pricing

Live £29/mo or £290/yr — save 17%

29/moStarting price

94%Cache hit rate

78%Earnings back

FreeTier available

EpisodicSemanticcryptographically sealedRedis-backed78% of margin backFree tier

How it works

Memory Engine in three steps

Production-grade and live on api.forcedream.ai. One Bearer token, zero extra setup.

01 🧠
Store

Every interaction is stored with episodic + semantic compression. Old memories compress automatically to stay within token budgets.
STORE /v1/memory/store
02 🔍
Recall

Top-k relevant memories are retrieved per task using semantic similarity. Context is injected automatically into routed calls.
GET /v1/memory/recall
03 💚
Earn

78% of margin on every API call flows back to you. Withdrawable via Stripe. cryptographically sealed for audit.
GET /v1/balance

What's included

Everything you need, nothing you don't

✓Episodic memory — every interaction stored verbatim

✓Semantic memory — compressed beliefs about each user

✓Auto-prioritisation — keeps what matters, drops noise

✓Context retrieval — top-k relevant memories per task

✓cryptographically sealed writes — immutable, auditable history

✓Cross-workspace sync — share memory across tenants

Quick start

api.forcedream.ai POST /v1/memory/store

FORCEDREAM OS · L501 LIVE

$ curl https://api.forcedream.ai/v1/memory/store \
  -H "Authorization: Bearer $KEY" \
  -d '{"user_id":"usr_123","content":"Prefers concise bullet answers","type":"episodic"}'

→ {"memory_id":"mem_9a3f","stored":true,"compressed":false}

$ curl https://api.forcedream.ai/v1/memory/recall \
  -H "Authorization: Bearer $KEY" \
  -d '{"user_id":"usr_123","current_task":"summarise Q3 results"}'

→ {"context":[{"content":"Prefers concise bullet answers","relevance":0.94}],"tokens_used":18}

Pricing

Simple, transparent pricing

78% of margin flows back to you on every call. No hidden fees. Free tier available.

Starter

£29

/mo · £290/yr

1M operations/month

10,000 users

Episodic + semantic

Standard retrieval

78% of margin

Start free →

How Memory Engine compares

Purpose-built for AI products. Not retrofitted from general-purpose tools.

Feature	ForceDream Memory	Pinecone	Weaviate	DIY Redis
Episodic memory	✓	—	—	—
Semantic compression	✓	—	—	—
WORM audit trail	✓	—	—	—
78% of margin back	✓	—	—	—
One-line API	✓	—	—	Manual
Free tier	✓	—	—	✓

FAQ

Frequently asked questions

What is episodic vs semantic memory?

Episodic memory stores specific events verbatim — e.g. "user asked to compare Q3 and Q4". Semantic memory compresses repeated patterns — e.g. "this user prefers bullet points". Both are indexed and recalled automatically.

Is memory shared across users?

No. Memory is partitioned by user_id. Shared workspace memories are available for team contexts via the workspace_id parameter.

Can I delete a user's memory for GDPR?

Yes. DELETE /v1/memory/{user_id} purges all records and writes a cryptographically sealed deletion receipt for compliance auditors.

How much context can I store per user?

No hard limit. The engine auto-compresses older memories to stay within token budgets. You set the retention window per API key.

How does the 78% of margin model work?

You earn 78p of every 100p of API revenue your product generates. The treasury (2.5%) funds the OS. The remainder covers infrastructure.