Ada documentation
Learn how Ada captures anything you share, understands what it is, and helps you take the next step from your phone, browser, or dashboard.
12
Content categories
25+
Shipping actions
3
Primary surfaces
Overview
What Ada Is
Ada is an AI share assistant. When you share a link, screenshot, text snippet, product, event, recipe, job post, or other useful content, Ada saves it, classifies it, extracts the important details, and suggests actions you can take next.
Ada solves the common problem of “I saved this, but now what?” Instead of leaving important links scattered across messages, notes, screenshots, and browser tabs, Ada turns shared content into organized items with useful follow-up actions.
At a high level, Ada captures content from the iOS share sheet, Chrome extension, or web dashboard; analyzes it with AI and structured rules; stores it in your account; and presents actions such as reminders, calendar events, notes, messages, research, and service-specific tasks.
Example: Event planning
You share a meetup page from Safari. Ada detects the date, location, and event category, then suggests adding it to your calendar, setting a reminder, or getting directions.
Example: Job search
You save a job posting from your browser. Ada extracts the company, title, deadline, and application context, then suggests creating a note, adding a follow-up, or starting an application flow.
Getting Started
Start In Five Steps
Get access
Ada is in public beta. Redeem an invite code at /beta, then create or sign in to your account.
Complete setup
Add basic preferences, choose how Ada should help, and optionally connect public X, Instagram, or LinkedIn profile data for personalization.
Install a capture surface
Use the iOS share extension from any app, install the Chrome extension, or add an item directly from the web dashboard.
Connect services
Link the services you want Ada to act in, such as Gmail, Google Calendar, Notion, Slack, GitHub, or Linear.
Save your first item
Share a link, screenshot, text snippet, product page, job post, recipe, or event. Ada classifies it and suggests the next useful actions.
First useful interaction
Share a real item you would normally forget: an event, restaurant, product, recipe, article, or job post. Open the dashboard, review the generated summary and category, then run one safe action such as creating a reminder or note.
Core Interface
How Users Interact With Ada
Share sheet
Example input
Share a restaurant link from Safari to Ada.
Expected output
Ada saves the item, categorizes it as Food & Dining, extracts location details, and suggests actions like find a restaurant, get directions, or save to Notion.
Dashboard item
Example input
Paste a job posting into the web dashboard.
Expected output
Ada summarizes the role, detects the company and deadline when available, and suggests actions like apply online, create a note, or add a calendar reminder.
Chat
Example input
Ask “What should I follow up on from my saved items?”
Expected output
Ada responds in conversation and may propose actions you can review before execution.
Searchable
Find items by text, category, urgency, saved date, and starred status.
Asynchronous
Some analysis and automations run in stages; pending items update when results are ready.
Reviewable
Actions with external effects show a review or confirmation step before completion.
Features
Core Capabilities
Capture
Save useful content from mobile, browser, or the dashboard.
What users can do
- Share links, text, screenshots, images, locations, contacts, events, and files.
- Add optional notes or intent when saving from supported surfaces.
- Find saved items later by search, category, starred status, urgency, or date.
Use cases
- Save a conference page before you forget the date.
- Capture a product page and track it later.
Ada works best when the shared content includes readable text, a public URL, or structured details such as dates, locations, prices, or contacts.
Classification & Extraction
Ada turns raw shared content into structured items.
What users can do
- Detect one of 12 categories such as Events, Food, Shopping, Travel, Jobs, Learning, Finance, and Social.
- Extract useful fields like dates, prices, contacts, locations, deadlines, recipes, and job details.
- Show pending items while analysis runs, then update the item when classification finishes.
Use cases
- A recipe link becomes ingredients and steps.
- A flight deal becomes travel context and price data.
AI extraction can be incomplete. Review important dates, prices, addresses, and recipients before acting on them.
Actions
Ada suggests practical next steps for each item.
What users can do
- Create reminders, calendar events, notes, contacts, Notion pages, Slack messages, emails, and issue tracker entries.
- Run browser-style actions such as research, price checks, availability checks, form fill support, and recipe extraction.
- Review editable actions before Ada sends messages, creates events, or changes external systems.
Use cases
- Turn an event page into a Google Calendar entry.
- Turn a GitHub idea into an issue with a drafted title and body.
Some actions require a connected service. High-impact actions show a confirmation or edit screen first.
Chat
Ask Ada questions and dispatch actions from a conversation.
What users can do
- Start and revisit conversations from the dashboard.
- Ask about saved items, follow-ups, or what to do next.
- Approve, edit, or reject proposed actions without leaving the chat flow.
Use cases
- Ask for a summary of recent job links.
- Ask Ada to draft an email based on a saved contact.
Chat responses are not a substitute for reviewing the source item. Treat suggestions as drafts until confirmed.
Profile & Memory
Ada uses your declared preferences and optional profile imports to personalize suggestions.
What users can do
- Set goals, preferred destinations, reminder frequency, content depth, interests, and work context.
- Import public profile signals from X, Instagram, or LinkedIn during onboarding or later from Profile.
- Apply or dismiss personalization suggestions so Ada can adapt without locking in guesses.
Use cases
- Prefer short summaries for articles.
- Tell Ada that job and learning items matter more than shopping items.
Inferred traits are suggestions, not facts. You control which suggestions become part of your profile.
Files & Rich Content
Ada can store and reason over richer shared content when supported by the capture surface.
What users can do
- Save screenshots and images with extracted text when available.
- Attach notes to items and keep generated artifacts with the source item.
- Review video or social content analysis when the item contains supported media metadata.
Use cases
- Save a screenshot of an event flyer.
- Keep extracted text from a social post with the source link.
Very large files, private URLs, or content behind logins may not be fully readable by Ada.
Integrations
Connected Tools
Integrations let Ada execute actions in the tools you already use. You connect services from the dashboard, and Ada only uses them for actions you approve or initiate.
Gmail
Send or draft emails from Ada-generated content.
Example: Send a follow-up email to a contact extracted from a saved page.
Google Calendar
Create calendar events from dates, deadlines, and plans.
Example: Add a workshop or interview to your calendar after review.
Notion
Save structured notes, recipes, jobs, research, or source summaries as Notion pages.
Example: Save an extracted recipe to a cooking database.
Slack
Send messages to channels or direct messages.
Example: Share a summarized link with a project channel.
GitHub
Create issues from technical notes, bugs, or project ideas.
Example: Turn a saved API bug report into a GitHub issue.
Linear
Create issues for product or engineering workflows.
Example: Capture a feature idea and send it to the right Linear team.
Browser automation
Research topics, check availability, compare prices, fill forms, and apply online with confirmation.
Example: Check whether a product is available or compare prices across pages.
Automations
Workflows & Monitoring
Workflows let Ada watch an item over time and trigger a follow-up action when a condition is met. The current workflow surface focuses on monitors and simple trigger rules.
A workflow starts with a monitored item, such as a price tracker or page monitor. You choose a field, an operator, a value, and the action Ada should create when the condition matches.
Price drop reminder
- Trigger
- When a tracked price becomes less than your target.
- Result
- Ada creates a reminder action so you can buy at the right time.
Page keyword summary
- Trigger
- When monitored page text contains a keyword such as “available”.
- Result
- Ada creates a summary or note action for the changed page.
Page change note
- Trigger
- When a monitored value changes from empty to a new value.
- Result
- Ada creates a note connected to the source item.
Workflow limits: current workflows are intentionally simple. They are best for price thresholds, page text changes, and single-step follow-ups. Complex branching, team approvals, and public webhooks are not part of the current user-facing workflow builder.
Advanced Concepts
How Ada Makes Decisions
Agent behavior
Ada chooses actions from a fixed action catalog. It should suggest actions that fit the item category instead of inventing unsupported actions.
Tool calling
When an action needs an outside service, Ada checks whether the relevant connector is available and routes the request through the connected provider.
Multi-step actions
Some actions can reuse earlier results. For example, a recipe extraction can feed a Notion save so the final page contains the recipe rather than only the source URL.
Realtime state
Items can appear as pending while analysis runs. The dashboard updates as classification, extraction, and action suggestions complete.
AI & Models
- Ada uses AI models to classify, summarize, extract fields, generate suggestions, personalize onboarding, and answer chat messages.
- Deterministic rules are used where they are more reliable, such as obvious URL or category patterns.
- Model output can be incomplete or wrong, especially for private pages, ambiguous screenshots, stale pages, or content with missing context.
Safety & Permissions
Controlled Automation
Review before external changes
Emails, calendar events, Slack messages, GitHub or Linear issues, reservations, applications, and form actions require review or confirmation before Ada completes them.
Direct actions are lower risk
Actions like saving an internal note, creating an Ada reminder, extracting a recipe, researching a topic, or checking availability may run directly.
Errors stay visible
If an action fails, Ada keeps the item and reports the failure so you can retry, edit missing fields, or connect the required service.
You remain the approver
Ada can draft, suggest, and prepare work, but important outputs should be checked before you rely on them.
Permission model
Ada can organize and suggest automatically. External actions require the relevant connector, and higher-impact actions require a confirmation, editable form, or both. If required fields are missing, Ada asks for more information instead of guessing silently.
Privacy & Security
Data Handling
- Your saved content, notes, actions, preferences, and connected-service metadata are used to operate Ada for your account.
- Connected-service tokens are stored encrypted and used only for actions you authorize.
- Third-party AI and infrastructure providers may process content to provide classification, extraction, chat, and automation.
- Ada does not sell personal information and does not use your content to train third-party AI models without separate consent.
- You can request account deletion by contacting privacy@tryada.app.
Authentication
Dashboard access requires an authenticated account. Beta access may require an invite code while the product is in public beta.
User isolation
Your saved items, actions, conversations, preferences, and connections are scoped to your account.
FAQ
Common Questions
How much does Ada cost?
Ada is free during public beta. The planned Pro price after general availability is $9.99 per month, with team pricing available by request.
Do I need the iOS app to use Ada?
The iOS share extension is the main capture experience, but you can also use the web dashboard and Chrome extension where available.
Can Ada act without asking me?
Ada can classify, summarize, and prepare suggestions automatically. Actions that send messages, create external records, or perform higher-impact browser tasks require review or confirmation.
Which integrations are supported?
Current user-facing integrations include Gmail, Google Calendar, Notion, Slack, GitHub, Linear, Chrome extension capture, and browser automation for supported web tasks.
What happens if Ada reads something incorrectly?
You can edit action fields before execution, ignore the suggestion, retry analysis where available, or add a manual note to the item.
Does Ada have a public API?
Ada is currently focused on the app, dashboard, extensions, and connected services. A public developer API is not part of the current product experience.
Can Ada access private pages?
Ada works best with content you explicitly share and public pages it can read. Private pages, paywalled content, and pages behind logins may have limited extraction.
Can I delete my data?
Yes. You can request account deletion by emailing privacy@tryada.app. Some legal or operational records may be retained only when required.
Need help?
Visit the help center for setup articles, or contact the Ada team if something is missing, confusing, or not working as expected.