Settings

Recall settings are accessed through the popup menu from the gear icon in the toolbar. The menu provides quick access to:

• Settings — Open Model Settings or Matter Settings
• Setup Wizard — Re-run the initial setup
• Logout — Sign out of your account (server mode only)

Connection Settings

Connection Settings determine how Recall operates — either entirely on your machine or connected to a shared server.

Setting	Description
Mode	Choose Entirely Local or Backend from Server
Server URL	The URL of the backend server (server mode only)
Test Connection	Verify the server is reachable and responding

Entirely Local

All AI models run on your machine. No server or internet connection is needed after initial setup. Your documents and data stay on your computer.

Backend from Server

Connect to a shared backend server. In this mode:

• Documents and matters are stored on the server, organized by organization
• AI processing can be handled by the server
• Multiple users in the same organization can collaborate
• You need to log in with your account credentials

To configure:

1. Click the gear icon in the toolbar
2. Select Settings
3. Under Connection, choose Backend from Server
4. Enter the server URL
5. Click Test Connection to verify
6. Save your changes

You can also access connection settings from the login screen.

Accounts & Organizations

When using server mode, Recall requires an account.

Registration

Registration is invite-gated — you need an invite PIN from your administrator to create an account.

1. On the login screen, click Register
2. Enter your Display Name, Email, Password, and Invite PIN
3. Optionally enter an Organization Name (or one is auto-created)
4. Click Register

Login

1. Enter your Email and Password
2. Click Log In

Organization Management

Each account belongs to an organization. Organizations keep data isolated between teams.

Setting	Description
Organization Name	Displayed in the UI; can be renamed from settings
Display Name	Your name as shown in the application

To rename your organization, access settings from the gear icon popup menu.

Logout

To sign out:

1. Click the gear icon in the toolbar
2. Select Logout from the popup menu

Model Settings

Model Settings configure AI models, search behavior, and Word integration for the entire application.

Word Integration

Configure how Recall communicates with Microsoft Word.

Setting	Description
Integration Mode	Choose Recommended (Apple Events/COM) or Advanced (Word Add-in)
Request Word Permission	Request macOS Automation permission (macOS only)
Test Connection	Tests the connection to Word
Certificate Status	Shows certificate trust status (Advanced mode only)

Learn more about Word Integration setup →

Semantic Search

Configure the AI models used for document search.

Setting	Description
Enable Semantic Search	Toggle semantic search on/off
Embedder	Choose which embedding model to use
Rebuild Index	Re-index all documents with the current embedder

Available embedders:

• ModernBERT — Good balance of quality and speed
• E5-Large-Instruct — Higher quality, more resource intensive
• all-MiniLM-L6-v2 — Fastest, lower quality
• Custom — Your own GGUF embedding model

Generation

Configure the AI model used for Clerk and text generation.

Setting	Description
Generator	Choose which generation model to use
Disabled	Turn off AI generation entirely

Available generators:

• Llama 3.1 8B Instruct — Bundled, good quality
• Phi-3.5 Mini Instruct — Bundled, faster
• ChatQA — Optional download, best for Q&A
• Custom — Your own GGUF generation model

Enhancement Features

Optional features that improve search and generation quality.

Setting	Description
Query Expansion	AI generates alternative queries to improve recall
Reranking	Re-ranks results using Core ML GTE Reranker
Retrieval LLM	Uses ChatQA for query planning (requires download)
Corpus Search	Enables backend knowledge corpus

Display

Visual and interface settings.

Setting	Description
Content Font Size	Adjust font size for generated text and pins (10–24 pt)

Custom Models (Advanced)

Add your own GGUF models for embedding or generation.

1. Place GGUF files in the models folder:

   ~/Library/Application Support/Recall/Models/
   

2. Click Refresh Custom Models
3. Select your model from the dropdown

Matter Settings

Matter Settings configure options specific to a single matter. Each matter can have different settings.

Basic Information

Setting	Description
Matter Name	Display name for the matter
Notes	Freeform notes about the matter

Citation Settings

Setting	Description
Citation Style	Federal or California citation format
Jurisdiction	Primary jurisdiction for this matter

Context Settings

Setting	Description
Practice Area	Legal practice area (optional)
Court	Specific court information (optional)
Document Type	Default document classification

These settings help Clerk provide more relevant assistance.

Accessing Settings

Settings are accessed through the gear icon popup menu in the toolbar.

Model Settings

1. Click the gear icon in the toolbar
2. Select Settings from the popup menu
3. Navigate to the Model Settings section

Matter Settings

1. Select a matter in the sidebar
2. Click the gear icon in the toolbar
3. Select Settings, then Matter Settings
4. Or right-click the matter and select Settings

Setup Wizard

Return to the initial setup:

1. Click the gear icon
2. Select Setup Wizard from the popup menu
3. Walk through the setup steps again

Status Indicators

The status bar shows the state of various system components:

Indicator	Meaning
Word	Microsoft Word connection status
Server	Backend server connection status
Indexing	Document indexing in progress
Corpus	Backend knowledge corpus status
Model	AI model loading status
Reranking	Reranker availability

Click an indicator for more details or to access related settings.

Resetting Settings

Reset Model Settings

To restore default model settings:

1. Open Model Settings
2. Scroll to the bottom
3. Click Reset to Defaults
4. Confirm the reset

Reset Matter Settings

To restore default matter settings:

1. Open Matter Settings for the matter
2. Click Reset to Defaults
3. Confirm the reset

Data Storage

Recall stores data in platform-specific locations:

macOS

Data	Location
Settings	~/Library/Preferences/
Documents & Matters	~/Library/Application Support/Recall/
Models	~/Library/Application Support/Recall/Models/
Indexes	~/Library/Application Support/Recall/Indexes/
Connection Settings	connection_settings.json in app data

Windows

Data	Location
Application Data	%APPDATA%\Recall\
Models	%APPDATA%\Recall\Models\

Server Mode

In server mode, documents and matters are stored on the server, organized per-organization. Each organization has isolated storage including its own database and document directories.