evaluate( )

Pattern

Jev Memory

Long-term agent memory with three Jev gates. Keep durable facts, recall what fits a token budget, and evict what went stale. Nothing is rewritten.

BeginnerSep 2026View the full pattern on desktop
Jev Memory preview

Problems solved

  • Keep only durable facts in agent memory
  • Recall the memories that fit a token budget for this turn
  • Evict memories that newer turns made stale

Use cases

  • Long-term memory for a chat agent
  • Filtering what an agent writes to memory
  • Compacting stale facts out of a memory store

AI SDK APIs

experimental_evaluate

Tags

jevevaluationmemorylong-term-memorybooleanscoreai-sdk

Requirements

Node.js
24+
Memory
256MB
Other
  • Vercel AI Gateway API key

Environment variables

External services

vercel-ai-gatewayupstash

Ready to use?

View the full interactive pattern with live preview and copy the source code.

View the pattern on desktop

Related patterns

Jev Table Query preview

Jev Table Query

Filter or sort a closed table with one Jev question per row. A column name never leaves the app.

jevevaluation+5
View pattern →
Jev Tool Permissions preview

Jev Tool Permissions

Prune an agent's tool list before the model sees it. Allow and deny rules run first; Jev asks one relevance boolean per remaining tool in a single call.

jevevaluation+7
View pattern →
Jev Email Tags preview

Jev Email Tags

Tag a closed inbox with one Jev boolean per tag. A stamp lands only when its probability clears a floor.

jevevaluation+4
View pattern →
Jev Action Gate preview

Jev Action Gate

One command at a time. It passes the gate only when a Jev boolean clears 0.90.

jevevaluation+4
View pattern →