--- title: "XERJ vs memvid: single-file agent memory, tested" canonical: "https://xerj.org/compare/xerj-vs-memvid" updated: "2026-10-31" source: "content/compare/xerj-vs-memvid.md" --- # XERJ vs memvid: one memory file or one local server? **Short answer** (updated 2026-10-31): XERJ or memvid? memvid keeps memory in one portable file with almost no dependencies and answered queries in 0.3 ms. In our test its default lexical search found 1 of 75 answers, because full questions must match every word. XERJ found 73.3% with BM25 and 94.7% with semantic recall. **Key number:** memvid packages data, embeddings, search structure and metadata into a single .mv2 file. Its lex feature is full-text search with BM25 ranking through Tantivy. (as of 2026-10-31; source: [github.com/memvid/memvid](https://github.com/memvid/memvid)) **TL;DR** memvid stores agent memory in one portable .mv2 file. XERJ runs memory as a local server you call over HTTP or MCP. memvid's queries were the fastest we measured. But its default search needs every word of a question to match, so natural questions mostly came back empty. ## Agent prompt ```text Act as a coding agent. Read https://xerj.org/llms.txt, install XERJ, start a local node with --embed-mode neural, store five facts about this project with POST /_memory/project, then recall each one twice, once with key words and once as a full question, using hybrid recall, and tell me which recalls missed. ``` ## Commands ### Command 1 Note: Install the single XERJ binary. ```sh curl -fsSL https://xerj.org/get | sh ``` ### Command 2 Note: Start a local node with the built-in MiniLM embedder. ```sh xerj --embed-mode neural ``` ### Command 3 Note: A full question works: hybrid recall ranks by any shared word and by meaning. ```sh curl -s -XPOST 'http://127.0.0.1:9200/_memory/project/_recall' -H 'content-type: application/json' -d '{"query":"What is the main database for this project?","hybrid":true,"k":5}' ``` *Figure: memvid details are from its README on 2026-10-31. In our run we used memvid-sdk 2.0.160 from PyPI.* ## Tested side by side Same 85 memories, same 75 questions. Top 5 is the share of questions where the right memory came back in the first five. | | memvid-sdk 2.0.160 | XERJ v1.0.0-rc.100 | | --- | --- | --- | | Top 5, default settings | 1.3% (lexical) | 73.3% (BM25) | | Top 5, questions that reuse the memory's words | 4% | 100% | | Top 5, semantic with MiniLM | Not tested | 94.7% | | New value first, changed facts | 0% | 13.3%; 93.3% with recency weight | | Median write | 67.3 ms | 11.3 ms | | Median query | 0.3 ms | 3.8 ms | | LLM calls per write | 0 | 0 | | What you install | 2 Python packages | One binary | | Storage | One .mv2 file | Index per namespace on a local node | | License | Apache 2.0 | Apache 2.0 | ## Why full questions missed Type a question or pick one. The left side keeps only lines that hold every word. The right side ranks any shared word, like BM25 in XERJ. *Figure: Runs in your browser and sends nothing. It is a small model of the two query styles, not memvid's code. The recorded memvid results are in probes/memvid_and_probe.json in benchmarks/agent-memory: full questions returned nothing, and the same key words found the memory.* If you use memvid, search with key words, not full questions, or strip question words first. ## When to pick memvid - You want memory as one file you can copy, ship or attach. - You want no server process at all. - Your agent searches with key words it knows are in the memory. - You want the smallest install. ## When to pick XERJ - Your agent asks in full sentences or in other words. - You want semantic or hybrid recall with no extra setup. - You want memory next to your logs, docs and code search. - Several agents need the same memory over HTTP or MCP. ## Limits on the XERJ side - XERJ queries were slower than memvid's in this test. - XERJ is a server, not a file you can pass around. - XERJ runs on one node with no replication. See the full [AI agent memory benchmark](/compare/ai-agent-memory-benchmark) for all seven projects. ## Try it with XERJ Use the Copy the prompt button and paste it into your coding agent. It stores five facts and checks both key word and full question recall. ## Sources checked All checked on 2026-10-31. - memvid: README in [memvid/memvid](https://github.com/memvid/memvid), commit e6bd9f7b. memvid-sdk 2.0.160 from PyPI. License Apache 2.0 from the GitHub API. - Test runs, the all words probe and per question results: [benchmarks/agent-memory](https://github.com/xerj-org/xerj/tree/main/benchmarks/agent-memory). ## FAQ ### Why did memvid score so low? Our questions are full sentences. memvid's lexical search returned nothing when a question held words that were not in the memory. The same key words with no question words did find the memory. We recorded both in a probe. ### Is that a fair test of memvid? It is a fair test of its default lexical search on natural questions, the same ones every system got. It is not a test of memvid's vector search, which we could not get working in the PyPI package. ### Why was vector search not tested? In memvid-sdk 2.0.160 the local embedding option returned error MV015, and a custom embedder stored no vectors, error MV011. So we report only lexical search. ### Where does memvid win? Queries took 0.3 ms median, the fastest in our test, and the SDK installs only 2 Python packages. The whole memory is one file you can copy. ### Where does XERJ win? Recall on natural questions, semantic and hybrid search with a built-in model, and writes: 11.3 ms median for XERJ BM25 against 67.3 ms for memvid. ## Evidence - Top 5: memvid lexical 1.3%, XERJ BM25 73.3%, XERJ semantic 94.7%. Same-words questions top 5: memvid 4%, XERJ 100%. New value first: memvid 0%, XERJ 13.3%, XERJ with recency weight 93.3%. — `benchmarks/agent-memory/out/summary.csv` - memvid lexical find() returned nothing for full questions and found the memory for the same key words. — `benchmarks/agent-memory/probes/memvid_and_probe.json` ## Related - [AI agent memory compared: real tests on 7 open-source projects](/compare/ai-agent-memory-benchmark) - [XERJ vs mcp-memory-service: two MCP memory servers, tested](/compare/xerj-vs-mcp-memory-service) - [XERJ vs ReMe (Alibaba Tongyi Lab): a tested comparison](/compare/xerj-vs-reme) - [XERJ vs Mem0: which memory fits your AI agent?](/compare/xerj-vs-mem0) - [BM25 vs vector search: which one should you use?](/answers/bm25-vs-vector-search) - [What's a simple way to give an agent long-term memory on my laptop without Qdrant?](/answers/agent-memory-without-postgres-qdrant) - [XERJ vs agentmemory: which memory for your coding agent?](/compare/xerj-vs-agentmemory) (same intent: agent memory) - [XERJ vs agent memory tools: Mem0, Zep, Letta and Cognee](/compare/xerj-vs-agent-memory-tools) (same intent: agent memory) - [Should agent memory use a knowledge graph?](/answers/knowledge-graphs-for-agent-memory) (same intent: agent memory)