AGENTMEMORY — PERSISTENT MEMORY FOR AI CODING AGENTS

GITHUB21.1KINSTALL

ZERO EXTERNAL DATABASES · v0.9.26<br>AGENTMEMORY<br>THE MEMORY LAYER YOUR CODING AGENT SHOULD HAVE HAD FROM DAY ONE. CAPTURE EVERY SESSION. RECALL IN MILLISECONDS. RUN ANYWHERE.<br>$npx @agentmemory/agentmemoryCLICK TO COPYSEE IT MOVESTAR21k

AS FEATURED IN<br>AlphaSignal180K technical subscribers<br>↗Linux Foundation backedPosition #19 · NEW 2026Live upvote count

95.2%<br>RETRIEVAL R@5 · LONGMEMEVAL-S

92%<br>FEWER INPUT TOKENS PER SESSION

53<br>MCP TOOLS

12<br>AUTOHOOKS

EXTERNAL DATABASES

1401<br>TESTS PASSING

01<br>HOOKS<br>12 AUTO-CAPTURE HOOKS PIPED INTO EVERY CODING AGENT. EVERY TOOL CALL, EVERY PROMPT, EVERY STOP BECOMES A COMPRESSED OBSERVATION.

02<br>RECALL<br>TRIPLE-STREAM RETRIEVAL — BM25 + VECTOR + KNOWLEDGE GRAPH. RERANKED ON DEVICE. P50 UNDER 20MS ON A LAPTOP.

03<br>CONSOLIDATE<br>HOURLY SWEEPS COMPRESS RAW OBSERVATIONS INTO SEMANTIC MEMORIES. DUPLICATES MERGED. STALE ROWS DECAYED. AUDIT ROW EMITTED EVERY DELETE.

12AUTO-HOOKS<br>CAPTURE EVERYTHING<br>Every PreToolUse, PostToolUse, SessionStart, Stop, and the rest fire into the memory pipeline without a line of glue code. Install the plugin, done.

53MCP TOOLS<br>NATIVE MCP SURFACE<br>memory_save, memory_recall, memory_smart_search, memory_sessions, governance, audit, export — full surface behind a single MCP server.

126REST ENDPOINTS<br>HTTP FIRST<br>Every MCP tool has a REST twin under /agentmemory/*. Curl it. Fetch it from the browser. Proxy it from your own agent.

BM25+ VECTOR + GRAPH<br>TRIPLE-STREAM RECALL<br>Hybrid retrieval pipes lexical, semantic, and relational scores through an on-device reranker. 95.2% R@5 on LongMemEval-S.

AUTOCONSOLIDATION<br>RAW → SEMANTIC<br>Hourly sweep compresses observations into semantic memories, merges duplicates, decays stale rows with retention scoring, and emits a batched audit row.

∞REPLAY<br>JSONL SESSION IMPORT<br>Point agentmemory at a Claude Code JSONL transcript and it rehydrates the full session — observations, tool uses, timeline — into the store.

GRAPHEXTRACTION<br>KNOWLEDGE GRAPH<br>Entities and relations extracted on compress. Query with /agentmemory/graph. Visualize in the viewer. Temporal edges supported.

MESHFEDERATION<br>PEER-TO-PEER SYNC<br>Register another agentmemory node, push / pull memories over authenticated HTTPS. Bearer-token required; no silent syncs.

MDOBSIDIAN EXPORT<br>YOUR NOTES, HYDRATED<br>Mirror memories to a sandboxed vault directory. Frontmatter-tagged markdown, ready for Obsidian's graph view.

5LLM PROVIDERS<br>BYO MODEL<br>Claude subscription (default, zero config), Anthropic API, Gemini, MiniMax, OpenRouter. Detected from env.

OTELOBSERVABILITY<br>TRACES + LOGS<br>iii-observability worker on by default. Exporter: memory for local, OTLP for Jaeger / Honeycomb / Tempo. Every operation produces a span.

0EXTERNAL DBs<br>ONE PROCESS<br>Runs as a single Node process. No Redis, Kafka, Postgres, Qdrant, Neo4j. State lives on disk as JSON. That's the whole stack.

VIEWER:3113 · LIVE OBSERVATION STREAMiii CONSOLE:3114 · ENGINE DASHBOARDSTATERAW KV BROWSER + JSON EDITORTRACESOTEL WATERFALL + FLAME<br>SHIP-WITH VIEWER · PORT 3113<br>The agentmemory server auto-starts a real-time viewer on port 3113. No install, no config. Everything the server sees, the viewer shows.<br>›LIVE OBSERVATION STREAM · EVERY HOOK AS IT FIRES<br>›SESSION EXPLORER · REPLAY ANY PAST SESSION<br>›MEMORY BROWSER · FILTER BY PROJECT / TYPE / CONFIDENCE<br>›KNOWLEDGE GRAPH VISUALIZATION · FORCE-DIRECTED<br>›HEALTH DASHBOARD · HEAP / RSS / EVENT LOOP LAG<br>$ open http://localhost:3113<br>SHIP-WITH VIEWER · PORT 3113

agentmemory@localhost:3111<br>REPLAYIDLE

AGENTMEMORYMEM0LETTACOGNEE<br>RETRIEVAL R@595.2%81.4%73.8%78.1%<br>EXTERNAL DEPS02 (Qdrant, Neo4j)1 (Postgres)1 (Neo4j)<br>REST ENDPOINTS121———<br>MCP TOOLS5112189<br>AUTO-HOOKS12000<br>NATIVE PLUGINS6———<br>OPEN SOURCEYES (APACHE-2.0)YESYESYES

BUILDERS USING AGENTMEMORY<br>IN THE WILD.<br>Verbatim from the Product Hunt launch thread. Each card links back to the source comment.<br>HOW THEY USE IT<br>Backfilled a month of Cursor transcripts<br>“I backfilled agent memory on my past month's Cursor agent transcripts. It was surprisingly accurate. Picked up on things that I moved away from.<br>Peter NeyraProduct Hunt ↗<br>Two weeks of production use<br>“Been using it for 2 weeks, and I definitely see improvements.<br>Pranav PrakashProduct Hunt ↗

WHAT THEY SAY<br>“Tackles one of the biggest pain points with coding agents: losing useful project context across sessions without bloating the context window.<br>Alper TayfurProduct Hunt ↗<br>“The focus on making memory actually useful for agents instead of just storing context endlessly.<br>Mia TaylorProduct Hunt ↗<br>“Memory often becomes just more noise over time. Agentmemory feels more intentional compared to a lot of tools in this space.<br>Thomas HallProduct Hunt ↗<br>“Tried it briefly — feels clean and easy to get started with.<br>Zoe AlexandraProduct Hunt ↗

FIRST-CLASS PLUGINClaude CodeFROM Anthropic

12 hooks + MCP + skills<br>NATIVE PLUGINCopilot CLIFROM GitHub

11 hooks + MCP · framed stdio<br>NATIVE PLUGINCodex CLIFROM OpenAI

6 hooks...