Best for
- Developers working on large projects who need persistent AI memory.
- Teams want to integrate IDE agents and GitHub Copilot with shared context.
- Engineers managing multiple repositories who need unified project intelligence.
M
Memory Journal MCP Server
Community
Persistent memory and context handoff across AI sessions, with deep GitHub workflow integration.
GitHub source repository ↗
70FMRS · B
Memory Journal MCP Server is a powerful, feature-rich tool for developers and teams that need persistent context and GitHub integration. It has extensive security features, but careful configuration is required.
Memory Journal MCP Server is an MCP server that provides project-level long-term memory for AI agents. It maintains persistent knowledge graphs to pass context seamlessly between AI sessions, eliminating "AI amnesia." Features include GitHub integration (issues, PRs, Actions, Kanban, milestones, etc.), dynamic project detection, hybrid search (keyword + semantic + date range), knowledge graph visualization, team collaboration (shared database + vector search + cross-project insights), Hush Protocol (async flags), and adaptive session briefings. It offers 70+ tools, 19 prompts, and 46 resources, plus a Code Mode that saves up to 90% tokens. Built with TypeScript, SQLite, and local embedding models, it is local-first and privacy-conscious.
Tools
create_entry
Create a new journal entry.
get_entry
Retrieve a single journal entry by ID.
get_recent_entries
Get the most recent journal entries.
update_entry
Update the content or metadata of a journal entry.
delete_entry
Soft-delete a journal entry.
list_tags
List all tags with usage counts.
merge_tags
Merge multiple tags into one.
tag_entries
Add or remove tags from entries.
search_entries
Full-text or fuzzy search across journal entries.
semantic_search
Perform vector search based on semantic similarity.
date_range_search
Search combining date range and keywords.
get_vector_stats
View vector index statistics.
get_statistics
Get statistical summary of the journal database.
get_cross_project_insights
Get cross-project analytics insights.
link_entries
Create a relationship link between two entries.
get_relationship_graph
Generate a knowledge graph in Mermaid format.
export_json
Export journal data as JSON.
import_json
Import journal data from JSON.
export_markdown
Export journal entries as Markdown.
import_markdown
Import journal data from Markdown files.
rebuild_vector_index
Rebuild the entire vector index.
add_to_vector_index
Add or update entries in the vector index.
get_github_issues
Get a list of GitHub issues for a repository.
get_github_issue
Get details of a single issue.
get_github_prs
Get a list of pull requests.
get_github_pr
Get details of a single PR.
get_github_context
Retrieve full repository context (including CI status, unreleased changes, etc.).
get_kanban_board
Get the contents of a GitHub Project Kanban board.
add_kanban_item
Add a new card to the Kanban board.
move_kanban_item
Move a Kanban card to another column.
delete_kanban_item
Delete a Kanban card.
get_github_milestones
Get a list of milestones with completion percentages.
get_github_milestone
Get details of a single milestone.
create_github_milestone
Create a new milestone.
update_github_milestone
Update milestone information.
delete_github_milestone
Delete a milestone.
get_repo_insights
Get repository traffic and analytics for the last 14 days.
create_github_issue_with_entry
Create an issue and automatically link it to a journal entry.
close_github_issue_with_entry
Close an issue and automatically record a journal entry.
get_github_actions
Get GitHub Actions workflow run history.
get_github_workflows
Get a list of workflow definitions.
backup_database
Backup the journal database.
list_backups
List all backup files.
restore_database
Restore the database from a backup.
cleanup_backups
Clean up old backup files.
team_create_entry
Create an entry in the team database.
team_get_entry
Get an entry from the team database.
team_update_entry
Update an entry in the team database.
team_delete_entry
Delete an entry from the team database.
team_search
Full-text search in the team database.
team_semantic_search
Semantic search in the team database.
team_get_stats
Get team database statistics.
team_link_entries
Link entries in the team database.
team_get_relationship_graph
Generate a relationship graph for team entries.
team_export_markdown
Export team data as Markdown.
team_import_markdown
Import team data from Markdown.
team_backup
Backup the team database.
team_restore
Restore the team database.
team_list_flags
List active Hush Protocol flags.
team_update_flag
Update flag metadata or status.
team_resolve_flag
Resolve or close a flag.
team_pass_flag
Create a new Hush Protocol flag.
team_get_flag_analytics
Get flag resolution velocity and workload statistics.
mj_execute_code
Execute JavaScript code in a sandboxed environment, accessing all mj APIs for multi-step operations.
Setup
- Install globally via npm:
npm install -g memory-journal-mcp - Add a configuration block under
mcpServersin your MCP client config (e.g., Claude Desktop, Cursor). - Set environment variables as needed, such as GITHUB_TOKEN, PROJECT_REGISTRY, and ALLOWED_IO_ROOTS.
- Alternatively, use Docker:
docker run -i --rm -v ./data:/app/data writenotenow/memory-journal-mcp:latest. - Restart your MCP client.
claude_desktop_config.json
{
"mcpServers": {
"memory-journal-mcp": {
"command": "memory-journal-mcp",
"env": {
"GITHUB_TOKEN": "ghp_your_token_here",
"PROJECT_REGISTRY": "{\"my-repo\":{\"path\":\"/path/to/your/git/repo\",\"project_number\":1}}",
"ALLOWED_IO_ROOTS": "/path/to/your/git/repo"
}
}
}
}
Fit and risk
Not for
- Users who need only a trivial, stateless tool without any configuration.
- Projects where persistent memory is explicitly not desired due to privacy concerns.
- Environments that do not permit local file storage or database files.
Required permissions
- Local filesystem access limited by the `ALLOWED_IO_ROOTS` environment variable, used for import/export and Code Mode.
- GitHub API access via the `GITHUB_TOKEN` to interact with issues, PRs, actions, etc.
- Network access for remote HTTP/SSE transport, fetching embedding models, and OAuth token validation.
- Read/write access to SQLite database files stored on the local machine.
Risks and side effects
- Misconfiguration of `ALLOWED_IO_ROOTS` could expose unintended file paths to the MCP tools.
- A compromised GITHUB_TOKEN could allow unauthorized access to your repositories.
- Code Mode executes sandboxed JavaScript, but there is always a risk if the sandbox is not perfectly isolated.
- HTTP transport without proper authentication (e.g., OAuth or token) could allow unauthorized access.
- Third-party embedding models may introduce supply chain risks.
Troubleshooting
- Verify that the MCP client is correctly configured with the right command and arguments.
- Check that all required environment variables, especially GITHUB_TOKEN and ALLOWED_IO_ROOTS, are set appropriately.
- If using multi-project support, ensure PROJECT_REGISTRY uses valid JSON and the paths are correct.
- If semantic search fails, confirm that the vector index is rebuilt (set AUTO_REBUILD_INDEX=true or run rebuild_vector_index).
- When using HTTP transport, ensure the server is reachable on the expected port and that authentication is properly enabled.
Use cases
Maintaining a long-term record of decisions and context for a project across multiple AI sessions.
Automatically generating session summaries and loading them at the start of each new session for continuity.
Managing GitHub issues, PRs, milestones, and Kanban boards while linking work to journal entries.
Enabling team-wide shared knowledge and asynchronous collaboration through Hush Protocol flags and vector search.
Visualizing project timelines, status reports, and relationship graphs.
Supported clients
Claude DesktopFull support
CursorFull support