Glossary

Memory, context, and intelligence —
translated.

Tempreon speaks the same language as the rest of the AI memory world — plus a few words of our own. Here's how the industry vocabulary maps to the product, so you can tell what's the same, what's different, and what's layered above.

Industry term → Tempreon term

Comparing Tempreon to Mem0, Zep, Cognee, Personal.ai, ChatGPT memory, or Claude memory? Here's the dictionary so every apples-to-apples check resolves cleanly. Each term has its own page with the full definition.

Industry termIn TempreonWhat it means
AI memoryKnowledge Vault + Core ImprintIn Tempreon, stored memory lives in the Knowledge Vault, and the identity layer above it — who you are, how you think, how you decide — is the Core Imprint. Memory is the floor. The layer above it is what makes output sound like you rather than like an AI.
Memory types (episodic, semantic, procedural)Instincts — Style, Process, JudgmentTempreon organizes learned patterns as Instincts, in three kinds: Style (how you write and communicate), Process (how you run work), and Judgment (how you make a call). These are learned from how you actually work rather than declared up front.
Context windowPersonalized contextTempreon assembles the context for each interaction rather than dumping an archive into it — drawing on stored memory plus the patterns learned about how you work, so the model starts informed instead of cold.
Knowledge graphYour Network + Knowledge Vault linksYour Network and the Knowledge Vault combine to form a graph of who and what matters in your work, mapped as you use your AI tools rather than assembled by hand.
Ontology / custom entity typesDomains + AgentsTempreon uses Domains to scope different areas of your work — personal, work, side projects — so the right memory is in the right place, and Agents to give each area a distinct way of operating.
Fact invalidation / temporal memorySupersessionTempreon calls this Supersession. When something you said is no longer true, the old version is marked superseded rather than deleted, so current truth and history are both available.
Memory layer / memory infrastructureThe Intelligence ForgeTempreon's full stack — memory plus the learning layer that compounds it into something that sounds like you — is the Intelligence Forge.
Context engineeringDynamic context assemblyTempreon does this automatically. Each interaction gets context drawn from your stored memory plus the patterns the system has learned about how you work, without you pasting a prompt.
Self-improving memoryTempering & HoningTempreon names these separately. Tempering is the passive learning from how you work; Honing is the active refinement when you correct the system. Both sharpen the intelligence layer over time.
Personalized retrievalPersonalized searchTempreon weights search results the way you would, so what matters to you gets found rather than merely what matches the words.
Portability / model-agnosticBridgesBridges are how Tempreon connects to AI tools — one identity layer across Claude, ChatGPT, Cursor and more, so your intelligence follows you to whichever tool you are working in.
System prompt / custom instructionsThe KernelTempreon's Kernel is the earned version of that file — versioned operating rules distilled from how you actually work, in your own words, portable to any LLM.
Memory migration / memory export-importMemory ImportMemory Import brings memory you have built in other AI tools into Tempreon, after which it travels to every tool you connect.
MCP server / custom connectorYour Tempreon Bridge endpointTempreon is delivered as a hosted MCP server. Add the Bridge URL once in any MCP-capable client and that tool gains your intelligence layer.
MCP client / MCP hostAny connected AI toolAny AI tool you connect to Tempreon is acting as the MCP client. The setup guides cover the major ones individually.

Common questions

If you landed here searching for one of these, the short answer is first and the full story is a click away.

See Tempreon, not just the words

The vocabulary is half the story. The product is the other half.