Back to Signal Feed
CodeTracked since May 19, 2026

Reroute MCP memory_recall to /agentmemory/search

This PR changes standalone MCP `memory_recall` proxy behavior to call `/agentmemory/search` instead of `/agentmemory/smart-search`, and passes through `format` and `token_budget` so `full`, `compact`, and `narrative` recall modes return correctly. It also updates local fallback handling to align with search-style `memory_recall` responses.

MCPmemory_recall/agentmemory/search/agentmemory/smart-search

What Happened

  • This PR changes standalone MCP `memory_recall` proxy behavior to call `/agentmemory/search` instead of `/agentmemory/smart-search`, and passes through `format` and `token_budget` so `full`, `compact`, and `narrative` recall modes return correctly. It also updates local fallback handling to align with search-style `memory_recall` responses.
  • This PR changes standalone MCP `memory_recall` proxy behavior to call `/agentmemory/search` instead of `/agentmemory/smart-search`, and passes through `format` and `token_budget` so `full`, `compact`, and `narrative` recall modes return correctly. It also updates local fallback handling to align with search-style `memory_recall` responses.
  • 1 evidence item attached for review.

What is Different

Before

Scattered source updates, isolated context, and manual follow-up across multiple feeds.

Now

Fixes the MCP memory recall path by correcting the proxy endpoint and forwarding `format`/`token_budget`, restoring consistent recall-mode behavior (`full`, `compact`, `narrative`) and preventing silent mode/response mismatch between local and proxy recall flows.

Why Track This

Why It Matters

Developers and operators using MCP `memory_recall` can now get recall output in the requested shape and budget, so downstream clients get stable, predictable behavior instead of unexpected fallback responses when invoking recall modes. The route now goes to `/agentmemory/search` with explicit format and budget parameters, so watch for integrations that still assume `/agentmemory/smart-search` semantics or depend on old fallback behavior, and validate that token budget handling remains correct for edge-case values.

Impact

Developers and operators using MCP `memory_recall` can now get recall output in the requested shape and budget, so downstream clients get stable, predictable behavior instead of unexpected fallback responses when invoking recall modes. The route now goes to `/agentmemory/search` with explicit format and budget parameters, so watch for integrations that still assume `/agentmemory/smart-search` semantics or depend on old fallback behavior, and validate that token budget handling remains correct for edge-case values.

What To Watch Next

  • Watch whether MCP becomes a repeated pattern.
  • Track follow-up changes around Model Context Protocol.
  • Compare future signals against this evidence trail.
  • Re-check risk flags: hardcoded_smart_search_clients, fallback_response_shape_regression.
Open Topic TimelineOpen Technical EventOpen Original Sourcehardcoded_smart_search_clients / fallback_response_shape_regression / token_budget_edge_cases

Supporting Evidence

GITHUB PULL REQUESTHigh Trust

rohitg00/agentmemory PR #532: fix(mcp): route memory_recall through full search

Route standalone MCP `memory_recall` proxy calls to `/agentmemory/search`, forward `format` and `token_budget`, and align local fallback `memory_recall` with search-style responses.