Install Anamnesis

1. Install the Extension

Install Anamnesis Cloud from the VS Code Marketplace:

• Publisher: sharadtech
• Extension ID: sharadtech.anamnesis
• Compatible with VS Code 1.90+ and Cursor

Or install from the command line:

code --install-extension sharadtech.anamnesis

For Cursor on macOS, use the bundled install script from the extension repository.

2. Configure Credentials

Open the Anamnesis activity bar and click the settings gear, or run Anamnesis: Open Settings.

Create an account at anamnesis.cloud, then copy your Client Id and Secret Key from Settings - View Credentials.

In Anamnesis Configuration you can Test Connection, Save, and Install MCP and Skill. The install button detects VS Code or Cursor and writes MCP plus the Anamnesis skill for that IDE.

3. Create Your First Graph

1. Open a workspace in VS Code or Cursor.
2. Right-click a project folder in the Explorer.
3. Select Anamnesis: Create Knowledge Graph.
4. Wait for extraction and upload to Anamnesis Cloud.
5. Open the project from the Projects tree in the activity bar.

For Java/Maven/AEM repos, add build output to exclude globs:

"anamnesis.excludeGlobs": ["target/", "dist/", "node_modules/"]

4. Enable AI Tools (Optional)

After saving Client Id and Secret Key, click Install MCP and Skill on the Settings page, or run Anamnesis: Enable AI Tools (MCP + Skill). This registers:

• An MCP server (anamnesis) for graph queries
• A localhost MCP server (anamnesis-credentials) for stored secrets
• An agent skill for the current IDE (Cursor or VS Code / GitHub Copilot)

To remove, run Anamnesis: Remove AI Tools (MCP + Skill).

5. Explore the Graph

• Table view - fast filtering for large graphs
• Graph view - Cytoscape layouts with community colors
• Click to source - open the original file at the correct line
• Node inspection - label, kind, community, neighbors

6. Project Prompts

Open Project Prompts in the activity bar (or Anamnesis: View Prompts on a folder).

• Title and Original Prompt (required)
• Prompt Parameters - optional values for placeholders (brand, locale, constraints)
• Skip verification via AI Model - store the original prompt as-is without AI improvement
• View - markdown preview of Original, AI Generated, and Parameters, each with Copy

7. Credentials for AI

Use the Credentials for AI view to store named key/value sets (database URLs, API tokens, environment secrets). Ciphertext is saved on Anamnesis Cloud; values are decrypted only inside the extension.

• Toolbar + opens Add Credentials; click a set to edit it
• Tag a set with one or more Anamnesis projects
• After MCP install, agents can call list_credential_sets (names and metadata) and get_credentials(name) (decrypt locally)