Eclaire User Guide

Work in progress: These docs are a work in progress. More content coming soon.

More resources: For setup, configuration, running, or contributing to Eclaire, visit our GitHub repository at https://github.com/eclaire-labs/eclaire.

Main Interface Overview

Dashboard

Main Interface

1. [1] Dashboard — Overview of all data and recent activity inside Eclaire.
2. [2] Processing — Shows background jobs for bookmark fetching, document analysis, OCR, image classification, and lets you retry anything that failed.
3. [3] Upload — For bulk uploading content including bookmarks, documents, notes, or photos.
4. [4] All — Consolidated view of every asset in the system.
5. [5] Smart Lists — Quick views for Pending, Due Now, Pinned, and Flagged content.
6. [6] Main Assets — Jump into Tasks, Notes, Bookmarks, Documents, or Photos for asset-specific data management.
7. [7] History — Timeline of recent changes, imports, and automations.
8. [8] Settings — Configure profile, account, AI preferences, notifications, and API keys.
9. [9] AI Assistant Panel — Opens the assistant dock on the right for contextual help.
10. [10] AI Assistant Full Screen — Expands the assistant into a dedicated page for long-form conversations.
11. [11] Notifications — Alerts for items due today or overdue; click to review actions.
12. [12] Theme — Toggles between light and dark mode.
13. [13] Account Menu — Access profile options, settings shortcuts, and sign-out.
14. [14] New Bookmark — Quick-create entry point available from any asset list.
15. [15] Search — Global search across all information include content, tags, etc.
16. [16] Filters — Narrow results by type, date, tags, or other facets.
17. [17] View & Sort — Switch between grid/list (plus gallery for photos) and adjust sort order.
18. [18] Asset Card — Item detail block with title, summary, source, and status indicators.
19. [19] Card Actions — Pin, flag, or ask the assistant about that specific item.
20. [20] Tags — AI-generated labels to help classify and discover content.
21. [21] Floating AI Assistant — Shortcut launcher for AI including chat, content upload and assigning tasks to AI.

Note: All asset data—and the insights extracted from it—are available to the assistant for answering questions, automation, and content creation.

Settings and Configuration

The Settings view includes six sections: Profile, Account, Assistant, Notifications, API Keys, About.

Profile

Take a moment to customize your profile so the AI can respond with the right tone and context. The display name is what the assistant uses when addressing you, and the bio shares background details or preferences the model should consider when generating answers. Timezone, country, and city fields help the assistant give accurate guidance whenever time or location is part of the conversation.

Account

Use this section to handle core account management tasks. You can change your password, delete all stored data, or permanently delete the account.

Assistant

Control how the AI behaves in conversations. Toggle streaming responses in the UI, and enable or disable “thinking” mode when you’re using a model that supports reasoning.

Notifications

Connect communication channels so Eclaire can send notifications out or to talk to the AI assistant directly. Telegram integration is available today, with more channels coming soon. Once connected, choose whether the channel delivers updates, accepts incoming messages, or both.

API Keys

Generate and manage API keys for external apps, devices, or scripts. Keys allow connecting clients and devices like an Apple Watch, iPhone, iPad using Apple Shortcuts. They also allow for custom integrations through the full Eclaire API.

About

System information including version information, timestamps and changelog.

AI Assistant

Eclaire’s uses multiple AI models and retrieval tools to answer questions, automate workflows, and produce new content across your workspace.

Capabilities

The system uses different AI models depending on the tasks at hand. It has models for processing information such as recognizing documents and images, coming up with tags, titles, descriptions, etc. It’s also using a model to interact with users, to answer questions, manage content and produce new content as requested. The quality and performance of the AI will depend largely on the models that have been selected by the admin when setting up the system. Larger, more powerful models will likely produce better results but require more resources to run.

Tools

Once content has been added to the system, AI has access to all that information and has a variety of tools it can use to answer requests from users. These tools allow the AI to search and filter through content, look at the content itself to answer questions, summarize, or use for various tasks. The AI also has the ability to create new content when asked to do so. For example, users may ask the AI to create a new note with information from photos and documents.

Conversations

The conversations with assistant can make reference to specific assets in the system (eg. user may want to discuss a specific photo), in which case the AI will focus only on that asset to provide more relevant answers. Or conversations may also more broadly ask about all notes, or documents with a certain tag, etc. Answers from the assistant may reference specific assets in which case clickable link previews will be returned to the users as part of the conversation. Conversation history is also stored and can be retrieved later on as needed.

---