OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Agent Memory & RAG

Mentedb

Mentedb is a Rust database engine for AI-agent memory. It stores conversational turns and retrieves relevant context for later prompts, using components including HNSW vector search, knowledge graphs, write-ahead logging, and speculative context preassembly.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Mentedb — OpenViST project profile in Agent Memory & RAG.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Mentedb when an agent application needs persistent memory across turns and a retrieval step that supplies relevant context to the model. It is also relevant when a team wants to choose between managed cloud access and running the database itself.

Who it is for

  • Developers building AI agents that need persistent conversation memory.
  • Teams integrating memory retrieval into Python, TypeScript, or Rust applications.
  • Builders seeking a managed cloud or self-hosted memory backend.

Documented uses

  • Save a user message and assistant response as an agent memory.
  • Retrieve relevant stored memories and return them as context for a later prompt.
  • Connect agent applications through an SDK, REST, or listed framework integrations.
  • Query a local memory data directory or a running MenteDB server with its CLI.
Before you choose

Requirements & limits

The project documents a managed cloud service requiring an API key and self-hosting through Docker or a Rust-installed server. It is marked beta, with APIs subject to change between minor versions.

  • Self-hosted use supports storage and recall without an LLM, but fact extraction requires an LLM provider and API key.
  • The cloud service is documented as having LLM configuration wired up for users.
  • Available documented integration routes include Python, TypeScript, and Rust SDKs, cloud REST, a CLI, and LangChain and CrewAI memory-provider packages.
  • Beta status and possible API changes may matter when adopting it as a production dependency.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Self-host MenteDB and use Python memory recall

Run the MenteDB server locally, then store a conversation turn and retrieve relevant memories with the Python SDK.

Before you start

  • Rust installation sufficient to use cargo install.
  • Python and pip; on Debian or Ubuntu, use a virtual environment or pipx rather than a system-wide pip install.
  • An LLM provider and API key are needed for fact extraction; storage and recall work without an LLM.

Steps

  1. Install the server with the command below, then start it with mentedb-server --data-dir ./data.

    Documented command
    cargo install mentedb-server

  2. Install the Python SDK with the command below (or use a virtual environment or pipx on Debian and Ubuntu).

    Documented command
    pip install mentedb

  3. Create a database with MenteDB("./my-agent-memory") and call process_turn with a user message, assistant response, and turn ID to store a turn.

  4. Call process_turn with a follow-up question, then use the returned result.context memories in the next prompt.

Expected outcome

A local MenteDB server and Python-backed memory workflow that stores a turn and returns recalled context for a follow-up.

Guide limitations

  • MenteDB is beta and APIs may change between minor versions.
  • Fact extraction requires MENTEDB_LLM_PROVIDER and MENTEDB_LLM_API_KEY; without an LLM, storage and recall are still available.
  • The server guide documents REST on port 6677 and built-in nightly maintenance.
Documentation 1 ↗Documentation 2 ↗
Source-based practical guide

Connect a cloud MenteDB memory workflow over REST

Submit an agent turn to MenteDB Cloud using its documented REST request.

Before you start

  • A MenteDB Cloud API key, set as MENTEDB_API_KEY.
  • A command-line environment with curl.

Steps

  1. Set MENTEDB_API_KEY to the key obtained from the MenteDB app.

  2. Send a POST request to the documented /v1/process_turn endpoint at https://api.mentedb.com, with bearer authorization and JSON content type.

  3. Include user_message, assistant_response, and turn_id in the JSON body.

Expected outcome

The documented cloud request submits an agent turn to MenteDB's memory workflow.

Guide limitations

  • This recipe documents submitting a turn; the source does not provide response handling steps here.
  • MenteDB is beta and APIs may change between minor versions.
Documentation 1 ↗Documentation 2 ↗
COMPARE OPTIONS

Compare options for Vector search

These projects document a common task and share a project role. This is a source-based comparison, not an execution benchmark or a guarantee of interchangeability.

ProjectDocumented purposePlatformAccess and requirements
MentedbSave a user message and assistant response as an agent memory.CLI, Self-hosted, ServerThe project documents a managed cloud service requiring an API key and self-hosting through Docker or a Rust-installed server. It is marked beta, with APIs subject to change between minor versions.
KektordbSave and retrieve semantic memories across agent sessions.Self-hosted, ServerThe README describes self-hosted use and displays an Apache 2.0 license badge.
LongbowStore and retrieve dense or sparse vectors with HNSW search.Server
Evidence and transparency

Sources & review

Official repository ↗Guide documentation ↗
Source record details
Source availability
Active
Sources checked
13 October 2026
Original website title
nambok/mentedb · Official homepage title · retained in its source language
How we review sourcesSuggest a correction