other / agent

AI Memory Vault

Set up an Obsidian vault as your AI's working memory with a build script and starter templates for Claude Code.

At a glance

Purpose
Set up an Obsidian vault as your AI's working memory with a build script and starter templates for Claude Code.
Runs in
Local agent with shell
Operating systems
Not stated
Needs
Runtime: Obsidian; Needs internet: Not stated; Needs accounts or API keys beyond your agent: Not stated
Files and data
Reads files: Yes; Writes files: Yes; Sends data to other services: Not stated
Tool type
Agent
Delivery
From GitHub
Version
Latest on the main branch
Access
Code on GitHub
Compatibility
According to the README, it runs on Claude Code, and any terminal AI that reads and writes files works with rough edges. It uses an Obsidian vault for the notes.
External costs
The README says the system is free to use, including commercially inside your own business under its licence, and offers voluntary support for the author. The README says a Pro plan is sufficient but does not name the vendor. Your coding agent or model provider may charge for usage.
Permissions
The system acts through your coding agent with the access you give it. The README gives a prompt asking Claude to set up the system from the remote repository. According to the README, the build prompt can install Obsidian, interview you and write vault folders, daily notes, a profile and recurring-task instructions. The README places CLAUDE.md in your working directory for loading each session and MEMORY.md under ~/.claude/projects/... to redirect native memory to the vault. The README's update procedure asks the agent to pull the repository and report changes, with the vault, notes and working-directory CLAUDE.md kept outside that checkout.
Usage terms
Creative Commons Attribution Share Alike 4.0 International, as stated in the repository. See its LICENSE file for the exact terms.

Code on GitHub

jaredrhod/ai-memory-vault

CC-BY-SA-4.0

700 stars185 forksLast pushed 31 Aug 2026 UTC

Give your AI a real, persistent memory. The open-source system plus templates that turn an Obsidian vault into your AI's working memory. No vector database, just markdown.

Topics: agent-memory ai ai-agents ai-memory anthropic claude claude-code claude-skills

Public GitHub data fetched · 0 days old

View on GitHub ↗

What it does

AI Memory Vault is a build prompt and template system for using an Obsidian vault as an AI assistant's working memory. The README describes storing notes outside the model; the GitHub description identifies Markdown rather than a vector database.

The README says to run ai-memory-vault.md inside Claude, which checks for Obsidian, installs it if absent, interviews you and builds a boot configuration, project folders, daily notes, a profile and recurring-task Jobs.

The README describes CLAUDE.md as the boot configuration, VAULT-INDEX.md as the vault's operating manual, DAILY-NOTE.md as a note template and MEMORY.md as a pointer from Claude Code's memory to the vault. The README describes AI Priming as reading selected notes before producing an output. The README says the boot template includes the author's agent persona, which you can retain, rename or replace.

Who it is for

People who use Claude Code, or another terminal AI that reads and writes files, and want it to keep its notes in an Obsidian vault.

Example tasks

Run the build script inside Claude to set up the vault and boot config.
Fill in the marked spots in a template, or ask your AI to interview you and fill it in.
Create daily notes from the daily-note template.
Set up Jobs so the AI handles a recurring task your way.
Prime your AI with a set of your notes before it writes an email.

Getting started

The README gives two routes. Inside a Claude Code session, you paste a one-line request that asks Claude to set it up from the repository's git address; the README quotes the exact sentence. Or you run the build script, ai-memory-vault.md, inside Claude. The templates have spots marked [FILL IN: ...] that you complete by hand or by asking your AI to fill them in. The repository also has a TROUBLESHOOTING.md file. The README has the exact commands.

Limitations

The README names Claude Code and says other terminal AIs that read and write files work with rough edges. The README says a Pro plan is sufficient without naming its vendor. The system uses Obsidian, and the README says the build prompt installs it if absent. The README's templates require your information. The README places the boot configuration in Claude Code's working directory outside the vault, where Claude Code loads it each session, and places the memory pointer in ~/.claude/projects/... rather than the vault. The repository can change at any time; check its README before you install.

Data practices

The README describes memory notes stored in your Obsidian vault outside the model, including your profile, projects and task-related knowledge. The README describes AI Priming as having the assistant read selected notes before producing an output. The README places the vault and working-directory CLAUDE.md outside the repository checkout used for updates. The README does not describe telemetry or the model provider's handling of those notes. What is sent where depends on your coding agent and model provider.

Community experience

Not rated yet

Sign in to share your experience.

Comments 0

No comments yet.

Sign in to comment

Report a listing concern