Skip to content

docs(memory): document agent.enabled opt-in (ref #12886) - #698

Open
adamscross04 wants to merge 1 commit into
LibreChat-AI:mainfrom
adamscross04:docs/memory-agent-enabled
Open

docs(memory): document agent.enabled opt-in (ref #12886)#698
adamscross04 wants to merge 1 commit into
LibreChat-AI:mainfrom
adamscross04:docs/memory-agent-enabled

Conversation

@adamscross04

Copy link
Copy Markdown

Summary

  • Documents the breaking change from danny-avila/LibreChat#12886 (merged 2026-05-01): memory.agent.enabled: true is now required for the automatic memory-extraction agent to run.
  • Configs that upgrade past #12886 with an agent block but no explicit enabled: true currently log a startup warning and silently skip automatic extraction. Manual memory operations still work.
  • The Memory Configuration docs page never mentioned the new property, so self-hosters (including ours) discovered the change via warning logs after memory silently stopped being written.

Changes

  • Added enabled: true to every agent: example in memory.mdx:
    • Top overview example
    • Agent-by-ID example
    • Custom Agent Configuration example
    • Complete Configuration Example
    • Custom Endpoints example
  • Documented enabled as an Agent Configuration Field, noting the opt-in default and linking back to PR #12886.
  • Added an info Callout under the top example that quotes the exact runtime warning users will see if the flag is missing.
  • Added a bullet to the Notes section summarising the opt-in behaviour.

Test plan

  • File is a single .mdx in content/docs/, no build tooling changes.
  • Preview the page renders cleanly (Callout + OptionTable used identically to other pages).
  • Confirm the wording in the warning-text codeblock matches the string in packages/data-schemas/src/app/memory.ts — this PR quotes it verbatim.

Context / motivation

Surfaced by production logs on one of our self-hosted instances (BT). First warning fired 2026-07-14 at 15:38 UTC after a deploy picked up a LibreChat version past the merge of #12886. Zero automatic memory writes have landed since. The referenced example in librechat.example.yaml was updated in that PR, but this MDX page was missed.

Documents the breaking change from PR #12886 (merged 2026-05-01):
`memory.agent.enabled: true` is now required for the automatic memory
extraction agent to run. Configs with an `agent` block but no explicit
`enabled: true` log a startup warning and silently skip extraction —
manual memory operations still work.

Changes:
- add `enabled: true` to every `agent:` example (top example, agent-by-id,
  custom agent, complete example, custom-endpoint example)
- document `enabled` as an Agent Configuration Field with the opt-in
  default and a link back to PR #12886
- add an info callout under the top-level example showing the exact
  warning text, and a bullet under Notes summarising the opt-in behaviour

Fixes documentation gap surfaced by BT prod: memory extraction stopped
silently after upgrading past PR #12886 because the docs still showed the
pre-change example.
@vercel

vercel Bot commented Jul 20, 2026

Copy link
Copy Markdown

@adamscross04 is attempting to deploy a commit to the LibreChat's projects Team on Vercel.

A member of the Team first needs to authorize it.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant