Skip to main content

Persistent vector memory and RAG for AI agents — MCP server with 150+ tools

Project description

Phorvec

Persistent vector memory for AI agents. One file per agent, MCP-native, local-first.

What is Phorvec?

Phorvec is an embedded vector database designed for multi-agent AI systems. Each AI agent gets its own isolated .avdb database file with vector search, key-value storage, and a knowledge graph — all accessible through the Model Context Protocol (MCP).

Key Features

  • One-file-per-agent architecture (.avdb files) — portable, isolated, no shared state
  • MCP-native — works with Claude Code, Cursor, Windsurf, Cline, and any MCP client
  • 150+ tools — context persistence, semantic search, team memory, conflict detection, skill library
  • Local-first — all data stays on your machine, zero cloud dependency
  • Hybrid search — vector similarity + BM25 keyword search with RRF fusion
  • Knowledge graph — entity and relationship tracking across sessions
  • Compression — PCA spine+delta encoding for storage efficiency

Quick Start

Install

pip install phorvec

Configure with Claude Code

Add to your Claude Code MCP settings:

{
  "mcpServers": {
    "phorvec": {
      "command": "phorvec",
      "args": []
    }
  }
}

Configure with Cursor / Windsurf

Add to your MCP configuration:

{
  "mcpServers": {
    "phorvec": {
      "command": "phorvec",
      "args": [],
      "transport": "stdio"
    }
  }
}

Use phorvec in your code (Python)

pip install phorvec
from phorvec import Memory
with Memory.open("./agent-memory", agent_id="assistant-1") as mem:
    mem.store("The user prefers metric units")
    for h in mem.recall("what units?", top_k=5):
        print(h.score, h.content)

pip install phorvec is the free (Community) library + the phorvec MCP CLI. Team/enterprise features (the phorvec.Team API) ship in the paid wheel from your account dashboard.

Architecture

Nexus (nexus.json)       — Central registry mapping agent IDs to file paths
    |
HandleManager            — LRU cache of open agent databases (default cap: 500)
    |
AgentDB (.avdb)          — Per-agent database with vector storage + KV store + graph

Configuration

Phorvec uses a TOML config file at ~/.phorvec/config.toml. The server runs with sensible defaults — no config file required.

Environment Variables

Variable Description
PHORVEC_LOG_LEVEL Log level (default: info)
PHORVEC_DATA_DIR Data directory (default: ~/.phorvec/data)
PHORVEC_STORAGE_BACKEND Storage backend: lancedb or lite

License

Proprietary. Copyright (c) 2026 Teressoft. All rights reserved.

The source code in this repository is closed and not licensed for redistribution. See LICENSE for the terms.

Use of the distributed phorvec binary (free Community tier or any paid tier) is governed by the End User License Agreement at https://phorvec.com/terms.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distributions

No source distribution files available for this release.See tutorial on generating distribution archives.

Built Distributions

If you're not sure about the file name format, learn more about wheel file names.

phorvec-0.1.2-cp38-abi3-win_amd64.whl (15.5 MB view details)

Uploaded CPython 3.8+Windows x86-64

phorvec-0.1.2-cp38-abi3-manylinux_2_28_x86_64.whl (16.1 MB view details)

Uploaded CPython 3.8+manylinux: glibc 2.28+ x86-64

phorvec-0.1.2-cp38-abi3-macosx_11_0_arm64.whl (14.4 MB view details)

Uploaded CPython 3.8+macOS 11.0+ ARM64

File details

Details for the file phorvec-0.1.2-cp38-abi3-win_amd64.whl.

File metadata

  • Download URL: phorvec-0.1.2-cp38-abi3-win_amd64.whl
  • Upload date:
  • Size: 15.5 MB
  • Tags: CPython 3.8+, Windows x86-64
  • Uploaded using Trusted Publishing? No
  • Uploaded via: maturin/1.14.1

File hashes

Hashes for phorvec-0.1.2-cp38-abi3-win_amd64.whl
Algorithm Hash digest
SHA256 0633f4ed1df7c350e547d16d44fc4b2343b0b1289e00ebd17aedb06239812cb8
MD5 7f1bb8854dc1ae8d0e9825b71f214e07
BLAKE2b-256 9a7f472a83cf6b92bd96c00d77d80385ff3385d33ff8ac0b1bc455d801ba1edc

See more details on using hashes here.

File details

Details for the file phorvec-0.1.2-cp38-abi3-manylinux_2_28_x86_64.whl.

File metadata

File hashes

Hashes for phorvec-0.1.2-cp38-abi3-manylinux_2_28_x86_64.whl
Algorithm Hash digest
SHA256 06523d4daf1dd9604e523a4503662615334763236515af9c9bd20514a3e0c430
MD5 932638f45072dd440f70db59ff7e894d
BLAKE2b-256 c524bbff26bc8ba92970eced9fe9183bff683c08b684ba5cb981b6299adce39a

See more details on using hashes here.

File details

Details for the file phorvec-0.1.2-cp38-abi3-macosx_11_0_arm64.whl.

File metadata

File hashes

Hashes for phorvec-0.1.2-cp38-abi3-macosx_11_0_arm64.whl
Algorithm Hash digest
SHA256 834b409569dfeab5c2c662ba6e7339358d19e297981d7d7e8401c42ac12e9a0f
MD5 83fa4b07f9064bf8e34e0fb2a6fd8f9a
BLAKE2b-256 64b6df19b2e5d1eed78f6e0152ea584edd48058fc6ddecbdbb3a4015f8213c2b

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page