Browse AI agents, workflows and apps published by the LiveContext community. Install one in a click, or start from it and make it yours.
163 published listings. The ones that ship an application are previewed below exactly as they run.
163 listings, most popular first.
Enter a stock ticker and get a real-time fundamentals + retail-sentiment dashboard. Combines Google Finance quotes, Google News headlines and aggregated community + social chatter into an AI-generated narrative with a bull-vs-bear verdict. Every check is logged for history.
Search job openings worldwide with rich filters (schedule, remote, salary, recency) and display matching positions as cards.
Every hour, fetches active price alerts, checks current prices via Polygon (stocks) and CoinGecko (crypto), and sends Telegram notifications when targets are hit.
Scrapes X/Twitter via Apify (kaitoeasyapi tweet scraper) twice a day on the keywords stored in a table, synthesizes via an AI agent, displays in a Twitter/X-style interface and logs the brief.
PokéAPI Pokédex: form (name/number + shiny) → slugify (core:code) → fetch live Pokémon data from the no-auth PokéAPI (core:http_request) → shape into {meta, pokemon} with stats/types/abilities + official artwork URL (core:code) → log lookup to table → retro Pokédex-device interface that re-fires the scan.
Look up a public profile by handle and fetch the bio, audience metrics and most recent published items via Apify. Results render in a split-pane dashboard (sidebar profile + masonry feed) with each search logged for history.
Let AI sort your Gmail into your own piles and draft your replies, up to 100 emails a run. Choose the emails with a Gmail search (unread from the last three days by default) and how many to read (10, 25, 50 or 100), or preview first without changing anything. Each email gets a pile, a one-line reason with a confidence and a summary, then the app acts in Gmail per pile: a label (Triage - <pile>, created the first time) and, if you choose, archive, mark as read or star, plus a reply draft saved in the email thread for the piles you pick (Urgent and Needs reply by default). In Settings: your piles one per line with what goes there, the Gmail actions of each pile, VIP and muted senders, the language, tone, length and signature of the drafts with a no-promise guard, the language and length of the summaries, and a list of newsletter senders you could unsubscribe from. Every sorted email becomes a card on your Triaged Emails board, urgent first, and the next run skips it. Nothing is deleted or sent. Needs a Gmail connection.
Weekly scrape of Meta Ad Library for tech automation/chatbot keywords, score by longevity and variants, store in table, dashboard.
TV show finder on the no-auth TVmaze API: form (query + max results) → build search URL (core:code) → fetch shows (core:http_request) → normalize into {meta, listings} with poster/rating/genres/network/status (core:code) → log search to table → marquee poster-grid interface that re-fires the search.
Compare AI-generated images side by side: OpenAI vs Google Gemini. Enter a prompt, both models generate simultaneously, results displayed in a comparison dashboard.
Turn your Typeform survey answers into insights your team acts on. The app reads the new responses of your form (the last 24 hours, 7, 30 or 90 days), rates each written answer positive, mixed, neutral or negative and sorts it into your own themes, then writes the insight in one line with the respondent's own words and a suggested follow-up for unhappy ones. Each insight becomes a page in your Notion database, the digest of the period is posted to your Slack channel (after a run with new feedback and every Monday), each negative response gets a follow-up task on your task board, and every insight lands on your Insights by theme board. Nothing is sent to a respondent. In Settings: your themes, what you sell, the language of the notes, the Notion database and its columns, the Slack channel, the Monday run and the tasks. Needs Typeform; Notion and Slack are optional.
Check the emails your colleagues report as phishing. Put them under a Gmail label ("Phishing report" by default): the app reads them, finds every link, including the one behind each button, and checks the 4 most suspicious new links per run with urlscan.io (a private scan by default) and VirusTotal (how many security vendors flag the domain and how old it is). Each link is judged malicious, suspicious or clean from those scans, the malicious domains are listed ready to copy into your mail or DNS filter, your Slack channel gets an alert, and every link lands on your Phishing board with what the email tries to do and a reply ready for the colleague who reported it. Links are always shown defanged, and your mailbox is only read: nothing is deleted, moved or sent. In Settings: scan privacy, domains you trust, the Slack channel and what it is alerted about, the language of your notes and your signature. Uses Gmail, urlscan.io and VirusTotal; Slack is optional.
Sort every new Zendesk ticket in one run. The app reads the tickets that arrived in the last 24 hours (or 3, 7 or 30 days), tells for each one whether it is a bug, a billing question, a how-to question, a feature request or something else, and how urgent it is. Bugs become Jira issues with a clear write-up for the engineers, billing tickets are posted in your Slack channel, how-to questions get a reply drafted from your own help facts, and each urgent ticket gets a task "Answer <requester>" on your task board with a first reply in it. Every ticket lands on your Tickets by category board with what the customer needs and what was done. Nothing is sent to a customer and no ticket is changed in Zendesk. In Settings: the Jira project and issue type, the Slack channel, tasks, what counts as urgent, the facts the replies may use, tone, language and signature. Needs Zendesk; Jira and Slack are optional.
Answer every new lead in seconds. Type or paste a lead, or let your website form send it to the app's webhook: the app tells what they want (a demo, a price, help with your product, another enquiry or spam) and writes a reply that shows you read their message, with your booking link for demo and pricing requests, saved as a draft in your Gmail (never sent). The contact is created or updated in HubSpot, found by email so nobody is added twice, your sales channel gets a Slack alert for demo and pricing requests, and the lead lands on your Lead pipeline board, spam in its own column. In Settings: what you sell, booking link, signature, where support requests go, tone, language, Gmail draft or preview only, HubSpot on or off, the Slack channel and which leads alert. Uses Gmail, HubSpot and Slack; each one can be switched off.
Know which new HubSpot companies to call first. The app takes every company added to your HubSpot in the last days (7 by default), reads its website, and rates how well it fits your ideal customer: hot, warm or cold, with a fit score and why. For the best leads it writes the angle to open the call, an opening line and two questions. The fit is written back to HubSpot in a Lead fit property (created the first time), every company lands on your Leads by fit board, and each hot lead gets a task "Call <company>" on your task board. Companies with too little to judge are flagged so you can add a website or a description. In Settings: companies per run, write to HubSpot or score only, the property and its values, call tasks, companies already scored and the language of the notes. Needs HubSpot.
Walk into every sales call prepared. The app reads the next 24 hours of your Google Calendar, keeps the meetings with people outside your company, looks each company up (its own pages and the news of the last 30 days) and writes a one-page brief per call: who they are, who you meet, the news worth knowing, talking points tied to what you sell, questions to ask, what to watch out for and an opening line. Each brief is saved as a Gmail draft addressed to you (nothing is sent) and as a card on your Briefs board (to prepare, to read, call done, follow-up sent); calls already briefed are skipped and past calls move to Call done by themselves. In Settings: how far ahead to look, recipients, your team's other domains, the calendar, the news window, news per brief, language, length and calls per run. Needs Google Calendar and Gmail.
Paste your contacts and say what you offer: each one gets a short, personal first email in your Gmail, saved as a draft for you to send (or sent right away, if you choose). Every weekday the app reads each conversation: replies are flagged with what the contact wants and a suggested answer, follow-ups are drafted after the number of days you set, bounces and finished sequences are closed. Every contact sits on a Campaign Contacts board: queued, sent, followed up, replied, closed. Needs your Gmail account.
Answer meeting requests without the back-and-forth. The app reads your unread Gmail emails (your Primary inbox from the last 7 days, or any Gmail search or label), keeps the ones that ask for a meeting, checks your Google Calendar for free times in your working hours, and drafts a reply proposing three of them in the sender's language, each time given to one person only. The replies wait in your Gmail drafts: nothing is sent. Every request becomes a card on your Meeting Requests board (new, proposed, booked) and moves to booked by itself once the meeting is in your calendar; a request already answered is never answered twice. In Settings: meeting length, how far ahead to look, working hours and days, lunch break, time zone, notice, gap around meetings, times per reply, emails per run, tone, language, signature and your meeting link. Needs Gmail and Google Calendar.
Type a job title and up to 4 locations: the app reads the matching Google Jobs postings, finds the pay each one shows (hourly, daily, monthly or yearly, in any currency) and puts the lowest, median and highest pay per location side by side, with the share of postings that show pay. Each run saves one row per location to your Benchmarks table. Needs a SerpApi account for the Google Jobs search.
Type a candidate, the role and their referees (name, email, how they know the candidate, one per line). Each referee's relationship is recognized (manager, peer, direct report, client), then tailored questions and a short email to copy are written for each one, in your language and tone. The checks go to a References board and a follow-up task per referee lands on your task board. Nothing is sent for you. No account needed.
Type the role you are hiring for, the city and the must-have skills. Talent Sourcer searches public LinkedIn profiles with a web search (no LinkedIn account needed), rates each one a strong, possible or weak fit from its title, snippet and country, and shows them on a radar with the strongest in the center. New profiles are saved to your Prospects table, on the Sourcing pipeline board (Found, Contacted, Replied, Interview, Passed), never twice, and each new strong fit gets a personal first LinkedIn message, ready in your task board with the profile link, for you to send yourself. In Settings, choose the country of the search, nice-to-have skills, the words that leave a profile out (recruiters by default), drafts for strong fits, close calls or none, your pitch, your signature, the language of the drafts, or a preview that saves nothing. Needs a SerpApi account for the search.
Type the role you are hiring for, its level and the competencies to assess (five common ones are filled in), then pick the interview length and the number of interviewers. For each competency you get 2 or 3 questions with follow-ups, in the format that fits it (a real past example, a realistic scenario or a short exercise), what a strong and a weak answer sound like, and a 1 to 4 scorecard. The agenda is planned by the minute, with who leads each block. The kit is saved to your Interview Kits table, one row per role and level that is updated when you build it again, and each interviewer gets a task on the task board to prepare their blocks. In Settings: interviewer names, a few lines about the role, 2 or 3 questions per competency, the language of the kit, and preview only. No account to connect.
Paste up to 10 company websites (one per line; bare domains and work emails are fine), say what you sell and who your ideal customer is, and get a lead board. Each company's own website is read (home page, and its about and pricing pages when it links to them) and sorted Hot, Warm or Cold for your ideal customer, with the scores so a close call shows as one. For each company: what it does, who it serves, the signals read on its site (business or consumer buyers, size, hiring, pricing, markets), the reason with a sentence quoted from the site, and a first-message angle to copy. Sites that cannot be read are reported, never guessed. In Settings, say what makes a lead hot and what disqualifies one, the pipeline status written for hot leads, the language of the notes, and whether to save to your Qualified Leads table, where a company already listed is flagged and never added twice. No account to connect.
Paste up to three YouTube links and, if you like, what you care about. Each video's transcript is read and turned into notes: title, channel, length and thumbnail, a TL;DR, chapters whose timestamps open the video at that moment, key takeaways, one or two quotes word for word, what it means for your focus, next steps when the video calls for them, and a watch, skim or skip verdict. A video without a transcript is reported with the reason, never summarized from nothing. Notes are saved to your Video Notes table with your tag, or shown as a preview only. Needs a SerpApi account.
Read your latest Google reviews, see what customers praise and complain about, and get a personal reply draft for each review, ready to copy. Type your business's name and city, or paste its Google Maps link (Share, then Copy link), and choose which reviews to read: newest, lowest rated or most relevant. Urgent complaints come first. In Settings, choose which ratings get a reply, how to sign, the tone, the language (the review's own or yours), where unhappy customers can reach you, and whether to save the drafts to your Review Replies table. Nothing is posted on Google for you. Needs a SerpApi connection for Google Maps.
Type a company name or website and get a research brief before you reach out: what they do and for whom, key facts, the recent news that is really about them (results about other things with a similar name are left out), signals, talking points, questions to ask, watch-outs and a fit score for your offer. Say what you sell and the brief is written around it. In Settings, choose your goal (sell to them, partner, invest, interview, or watch a competitor), the country and period of the news, the language of the brief, and whether to save it to your Company Briefs table. Needs a SerpApi connection for the Google and Google News searches.
Paste up to five links and get, for each page, a TL;DR, key points, exact quotes, what it means for you, next steps and a verdict: read, skim or skip. Say what matters to you (for example pricing changes, or what a small agency should know) and every summary looks for it. In Settings, choose the summary length (3, 5 or 8 key points), the language (the page's own or yours) and the tag saved with each page in your Reading List table, or preview without saving. Pages already in your list are flagged, links that cannot be read are reported. No account to connect.
Any workflow, agent or application you build in LiveContext can be published to the marketplace, with its screens and tables, for others to install in their own workspace.
