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

Caura

Caura is an open-source Python memory service for AI agent fleets. It gives agents a shared place to write and retrieve knowledge, with MCP integration, knowledge-graph support, scoped visibility, trust policies and audit trails.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Caura — 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 Caura when agents need to retain and retrieve knowledge across interactions, or when a team wants agents in a fleet to reuse lessons under defined sharing scopes. The documented standalone setup is a way to try keyword-based write and recall locally; semantic recall requires an embedding provider.

Who it is for

  • Teams connecting multiple AI agents to shared operational knowledge.
  • Developers evaluating a local memory service for agent write-and-recall workflows.
  • Organizations that need controls over which agents can access shared memories.

Documented uses

  • Store information learned during agent interactions.
  • Retrieve relevant memories for later agent tasks.
  • Share a team-scoped lesson between agents in the same fleet.
  • Apply visibility scopes and trust rules to memory sharing.
  • Keep audit trails for memory activity.
Before you choose

Requirements & limits

The README documents a local standalone mode that does not require an API key or signup. The repository is licensed Apache-2.0.

  • Keyless standalone recall is keyword-based; semantic paraphrase search requires an embedding provider.
  • Visibility can be scoped to an agent, team or organization; organization-level recall is subject to a trust ladder.
  • The documented local stack includes PostgreSQL with pgvector, Redis, an API and a scheduler.
  • The project recommends agent-scoped credentials for production clients.
  • The repository identifies Caura as the current name and says supported MemClaw tool calls and environment aliases continue to work.

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

Set up Caura locally and try keyword recall

Start the documented standalone service, write a memory and search for it by keyword.

Before you start

  • Docker Compose
  • A local copy of the Caura repository

Steps

  1. Clone the repository and change into its directory.

  2. Copy .env.example to .env and set IS_STANDALONE=true.

  3. Start the services with docker compose up -d --wait.

  4. Use the documented local API requests to write a memory and search for matching keywords.

Expected outcome

A single-tenant local Caura instance can store a memory and return a keyword search result without an API key.

Guide limitations

  • Standalone mode bypasses authentication and is single-tenant.
  • Without an embedding provider, memories are stored without embeddings and recall is by keyword; semantic paraphrase search requires configuring an embedding provider.
  • This setup guide does not establish production readiness or workload performance.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
9 October 2026
Original website title
caura-ai/caura · Official homepage title · retained in its source language
How we review sourcesSuggest a correction