需要与我们合作推广您的 GitHub 仓库或 Hugging Face 页面或产品发布或网络研讨会等吗？请联系我们： https://forms.gle/wbash1wF6efRj8G58
Michal Sutter 是一位数据科学专业人士，拥有帕多瓦大学的数据科学硕士学位。凭借在统计分析、机器学习和数据工程方面的扎实基础，Michal 擅长将复杂的数据集转化为可操作的洞察。
Coding agents spend a large share of their tool budget on search. When the target is a known symbol, ripgrep：https://github.com/BurntSushi/ripgrep answers it exactly. When the target is a behavior described in plain language, keyword matching often misses, and the agent falls back to guessing terms, reading whole files, and assembling context by hand. Each of those detours costs tool calls, tokens, and wall-clock time.
The Qwen Developer team announced：https://x.com/QwenDevs/status/2095157452904018263 zg (zvec-grep), an open-source local-first search layer that puts semantic search, BM25, and ripgrep behind one interface for both humans and agents. The code ships under the zvec-ai GitHub organization：https://github.com/zvec-ai/zvec-grep with an Apache 2.0 license.
Is it deployable? Yes, today. It installs from npm as @zvec/zvec-grep ：https://www.npmjs.com/package/@zvec/zvec-grep, requires Node.js 22 or newer on macOS, Linux, or Windows, needs no GPU with the default model, and the Apache 2.0 license permits commercial use.
zg indexes a workspace once and then exposes several ways to query it. The retrieval pipeline docs：https://github.com/zvec-ai/zvec-grep/blob/main/docs/04-pipeline.md define four routes: a hybrid default that combines intent with lexical anchors, --fts for BM25-ranked exact terms, --vector for conceptual similarity with no lexical ranking, and --rg for exhaustive literal or regex matching. The first three read the index. The --rg route needs no index at all, which matters when a repository has not been indexed yet.
An anonymous workspace index lives in /.zvec-grep/ . Both .git and .zvec-grep are always excluded, along with common dependency, build, cache, and log directories, plus anything the repository’s own ignore rules exclude. Re-running zg index updates incrementally; changing the embedding model requires an explicit --rebuild because vector spaces from different models are incompatible even at matching dimensions.
Indexed results report a freshness state of fresh or possibly_stale , so an agent can act on a good-enough result instead of running a status preflight first.
zg install detects Codex, Claude Code, Cursor, and OpenCode on the machine and wires up the local MCP integration. The server speaks Streamable HTTP MCP on a loopback-only endpoint at http://127.0.0.1:7999/mcp , with optional bearer authentication.
The design decision worth noting is restraint. Per the MCP guide：https://github.com/zvec-ai/zvec-grep/blob/main/docs/03-mcp.md, the default agent toolset exposes exactly two tools: zvec_grep_search for when the intent is known but the exact string is not, and zvec_grep_rg for when a symbol, path, or regex is known. Index lifecycle stays with the CLI. A six-tool compatibility set that adds index create, drop, status, and server status exists but is opt-in through zg server on --mcp-toolset full , and the docs state that an agent must never silently create, rebuild, or delete a persistent index.
Output is shaped for context economy. Results come back grouped by file with line spans, and indexed source previews are omitted by default unless requested. zg also rejects output-changing ripgrep flags such as --json , --count , -l , and --vimgrep so the compact result format holds.
The embedding catalog：https://github.com/zvec-ai/zvec-grep/blob/main/docs/07-embedding.md currently documents ten local models and three remote Qwen endpoints. The quickstart default, local/potion-code-16m-v2 , is a Model2Vec static model with a 256-dimension output and an 8,192-token input limit; because it uses static vector lookup, selecting a GPU does not speed it up. Heavier local options include jina-embeddings-v2-base-code , embeddinggemma-300m , and qwen3-embedding-0.6b . Remote options run to qwen/qwen3.7-text-embedding at 128,000 input tokens and the multimodal qwen/qwen3-vl-embedding .
Remote use is gated. Configuring a provider credential does not authorize data transfer; that requires either --allow-remote for a single command or a signed workspace grant via zg auth grant , revocable with zg auth revoke . The launch post cites eleven on-device models against ten in the current docs, a small discrepancy worth flagging.
The evaluation numbers appear in the launch post, not in the repository, where the benchmarks section is still a placeholder. Both runs were paired A/B tests holding agent, model, prompt, runtime, and task constraints fixed, with the zg condition adding only a prebuilt index, MCP tools, and usage guidance. Index build cost is excluded from the tables.
On a 20-question SWE-QA-Bench sample, zg cut tool calls by more than half and input tokens by nearly half while raising the Judge score by 1.50 points. On an 80-question BrowseComp-Plus sample, accuracy moved from 98.67% to 99.00% while input tokens fell 37.56%, tool calls 43.52%, and agent time 38.58%. Separately, indexing the Django repository (3,457 files) is reported to finish in under 30 seconds on an Apple M4 Pro.
Sample sizes of 20 and 80 questions are small, and the reported reductions come from the vendor’s own runs, so independent replication is the obvious next step.
Need to partner with us for promoting your GitHub Repo OR Hugging Face Page OR Product Release OR Webinar etc.? Connect with us ：https://forms.gle/wbash1wF6efRj8G58
Michal Sutter is a data science professional with a Master of Science in Data Science from the University of Padova. With a solid foundation in statistical analysis, machine learning, and data engineering, Michal excels at transforming complex datasets into actionable insights.
扫码直达文章详情