Discover the best tweets about Perplexity AI, including search workflows, product launches, research techniques, and user comparisons. Updated every two weeks.
Practical Perplexity search and research workflows, meaningful product updates, and evidence-based comparisons.
Original Xholic analysis
Across 50 posts from 40 creators, Perplexity conversation is overwhelmingly supportive (92% SUPPORTIVE / POSITIVE) and media-heavy (86% of posts include media). Perplexity Computer dominates by volume at 44% of posts, but the highest-scoring themes are developer platform (median all-time score 86.627) and open-source releases (median 79.581), far above the overall median of 9.05. Research-workflow guides also outperform, with a theme median of 29.41. Critical voices (8%) flag hallucination and degraded results (2079612494830518601), slower news freshness than rivals (2037281485208555619), monetization pressure (2040556040610046006), and the Amazon–Comet agent access dispute (2050264413438116
92% of posts
All-time engagement
100% of posts
Published in 90 days
Conversation map
Posts about Perplexity Computer (the autonomous agent product): capability overviews, real-world task walkthroughs (competitor research, business plans, document cleanup, web app building, marketing workflows), scheduling/parallel agents,成本
44%
Prompt libraries, researcher-style prompting guides, Deep Research/Pro Search/Labs/Spaces usage, structured research workflows, and citation-driven research practices
12%
Perplexity API as a full-stack model-agnostic agent platform, Search SDK and cookbooks, Search-as-Code architecture, Decisions API, and agent-building infrastructure
10%
Head-to-head evaluations of Perplexity Computer vs. OpenClaw, Claude Cowork/Claude Code, and other agentic tools, with tradeoffs around control, hosting, ecosystems, and cost
8%
Valuations, revenue growth, NVIDIA investment, Aravind Srinivas narratives and interviews, strategy (local inference, agents over search), and monetization pressure signals (credit costs, upselling to Computer/Deep Research)
8%
Comet's side-panel co-browsing, in-context assistance, background task automation, and browser-agent behavior including the Amazon dispute
6%
Model-agnostic routing across GPT/Claude/Gemini/Sonar, Model Council multi-model comparisons, new model additions (e.g., Claude Opus 5.5), and Effort orchestrator behavior
6%
Perplexity's open-source contributions (pplx-decider, pplx-embed, Lily, PII-Tracer, WANDR, Numbat, Bumblebee) and dependencies like the Browser Use library
6%
Tone and stance
Performance benchmark
Posts with media make up 86% of this collection. Their median all-time score is 9.25, compared with 2.88 for text-only posts.
Format mix
Consensus and debate
What performs
Statistical standouts
Creator landscape
The five most represented creators account for 20% of the selected posts.
1. Arsalan
@AIwithArsalan
2 posts
2. Aravind Srinivas
@AravSrinivas
2 posts
3. Ole Lehmann
@itsolelehmann
2 posts
4. Julian Goldie SEO
@JulianGoldieSEO
2 posts
5. Justin Thomas
@JustinThomasAI
2 posts
6. Perplexity
@perplexity_ai
2 posts
The top theme by volume (44% of posts) centers on Perplexity Computer agentic workflows: the official Perplexity account (authorName 'Perplexity') announced mobile availability with cross-device sync, iOS now and Android coming soon (2032494752642568417); Brett Caughran described using a validation space in Computer for final checks and debugging of an Excel Uber model, calling agentic workspaces a big unlock while,
The developer-platform theme has a median all-time score of 86.627 and open-source releases a median of 79.581, both far above the overall median of 9.05. Highlights include the official Perplexity account positioning the API as a full-stack, model-agnostic agent platform that replaces your model provider, search layer, and embeddings (2031828396435771563); the Perplexity Developers account launching the Decisions
The highest-scoring tweet overall (all-time score 321.58, 35.53x the median) comes from Magnus Müller (username mamagnus00), who showed Perplexity is built on the open-source Browser Use library and traced a random 'capital of France' search bug to a hardcoded sanity-check prompt in browser_use agent code that was never disabled (2042339700082610345). Separately, Aravind Srinivas listed recent Perplexity open-source
Since the previous snapshot
Themes, sentiment, stance, and post format are classified per tweet. All counts, shares, medians, creator concentration, freshness, and performance comparisons are then calculated directly from the published snapshot.
Xholic's all-time score compares engagement while accounting for reach, post age, and creator consistency. It is used for relative comparisons within this collection.
This report analyzes the exact 50-post snapshot shown below. AI identifies editorial categories and drafts explanations; all statistics are calculated from the snapshot, and every narrative claim is checked against cited posts before publication.
Best Perplexity tweets
Ranked 01–50
@mamagnus00 ·
I just realized that Perplexity is built on Browser Use open-source library. Last April, Perplexity users kept reporting that it was randomly searching for “capital of France” and answering “Paris” for unrelated prompts. That exact prompt, “What is the capital of France?”, is hardcoded in Browser Use. We used it as a sanity check in _verify_llm_connection: every time an Agent() was instantiated, it sent that prompt to the LLM. You can disable that but they forgot. Honestly, if they'd just told us, I'd have happily shown them how to integrate it properly. Feels like with Manus. Commit in browser_use: browser_use/agent/service.py lines 1272–1296 at commit 3f4c918a
@perplexity_ai ·
Perplexity Computer is now on mobile. Start any task on any device. Manage Computer from your phone or desktop with cross-device synchronization. Available now for iOS in the Perplexity app. Coming soon to Android.
Watch video
@FundamentEdge ·
It's really remarkable how fast AI tools for Excel have evolved. Even three months ago I found them almost completely unusable. Today, I was able to update my Uber model for the last four quarters in a fraction of the time, accurately, even when I consider the time I spent de-bugging and validating the key inputs. The three big unlocks for me were creating my own skills files, which are recipe cards encoding an incredibly detailed dissection of every step of the financial modeling process (put together in an 86 page document then crafted into six distinct modeling skills...unfortunately, I won't be sharing this at this time, but will consider in the future), connecting the Daloopa MCP to Claude in Claude Excel for accurate data, and creating a validation space in Perplexity Computer to do final checks and de-bugging. (I am not sponsored by either Daloopa or Perplexity, or any vendor for that matter) Obviously this AI augmented process is only valuable to the extent that it is 98%+ accurate and 100%+ accurate on critical metrics. Validation has to be a systematic process blending coding tools and human validation checklists (i.e. hand checking key model variables and understanding where in the model there is tolerance for mistakes, and where there isn't). But the ability of new LLMs to read & analyze models (particularly GPT 5.4) and the rise of Agentic Workspaces like Perplexity Computer to route tasks to the right LLMs seems to be resulting in big progress here. Really exciting stuff. I have been a huge skeptic here...Excel-based models are the foundation of institutional decision making, and they are no place for AI slop. With the technology improving, particularly workflows around systematic validation, that skepticism is melting.
@jspeiser ·
I've been using all three AI agents. OpenClaw, Perplexity Computer, and Claude Cowork. Here's what I've found so far: OpenClaw is the most powerful if you're technical, love tinkering and have a ton of time to mess around. Open source, self-hosted, you pick your models. You can make it do basically anything. But you're also the one building and maintaining all of it (as its always breaking). Claude Cowork feels the most like having an actual coworker (for 1 specific task). It runs on your Mac, works with your local files, and the plugins for legal, finance, and HR are legit. Downside is it only uses Anthropic models and your laptop has to stay open. Perplexity Computer is the easiest one to actually get work done (imo) with tons of models running behind the scenes, plugs right into Gmail, Slack, Notion, and it runs in the cloud so it keeps going while you sleep. Tradeoff is you're living in their ecosystem. (oh and their local browser control needs major work, it's unreliable) so: Want full control? OpenClaw. Want a desktop co-pilot? Cowork. Want to hand off real tasks and walk away? Perplexity Computer. These tools are changing weekly though, im just trying to keep up. What did I miss?
@perplexity_ai ·
The Perplexity API platform is now a full-stack, model-agnostic API platform for building agents. It replaces your model provider, search layer, and embeddings, built on the same infrastructure that powers Perplexity.
@perplexitydevs ·
Introducing the Perplexity Decisions API. It's powered by pplx-decider-v1-27b, our multimodal decision model trained to output a probability distribution over a fixed set of answers instead of text. It costs $0.04/million input tokens and scores 85.71% across benchmarks.
@BradoCapital ·
I'm pleased to announce we've expanded our partnership between @Perplexity_ai and @Fiscal_ai. Fiscal AI's Data API now gives investors on Perplexity and Perplexity Computer accurate, source-backed financial data when performing research or vibing customized dashboards.
@AskPerplexity ·
We just shipped several upgrades to building web apps with Perplexity Computer. New projects are now automatically tested using Playwright. It navigates your app like a real user to identify and fix bugs before you ever see them. Build, test, fix, ship. All in one place.
Watch video
@AravSrinivas ·
a few open source contributions from perplexity recently: •pplx-decider-v1-27b: SoTA multimodal decision model. 85.7% average across 11 benchmarks, ahead of Jev. •pplx-embed-v2-context-9b-preview: SoTA contextual embeddings, best on ConTEB and turbopuffer context-bench. •Lily: local inference engine for Apple silicon. rust plus custom metal kernels, no pytorch or mlx. 1.23x faster prefill and 1.35x faster decode than MLX-LM on an M5 Max •PII-Tracer: 0.6B on-device PII classifier that decides when a hybrid compute task stays on your Mac. beats OpenAI’s Privacy Filter on all 5 public benchmarks. also released with the PII-TRACE benchmark: 13k conversations in 13 languages •WANDR: benchmark for wide and deep research agents. 500 tasks needing 170k source-backed records •Numbat: agent detection and response for laptops and workstations. 52 rules, single go binary for macOS, linux and windows •Bumblebee: read-only supply chain scanner for dev machines. covers packages, MCP configs, and editor and browser extensions. a lot more open source contributions coming soon!
@EXM7777 ·
your workflow's biggest weakness is research... built-in web search on Claude, ChatGPT, Gemini... it's all surface level i got tired of this so i built a skill that fixes it: > reads the full conversation and project context > generates a research plan with multiple angles > fires parallel prompts through Perplexity's sonar deep research > processes all outputs and merges them into one clean markdown file i trigger it manually because not everything needs deep research... but when it does, the difference is night and day the whole is relatively cheap for the bump in quality you're getting
@sakhil_ai ·
8 Perplexity Prompts For Research🟡 Turn Perplexity into a world-class research analyst. 1. Market Landscape Snapshot Analyse the current market landscape for [INDUSTRY / TOPIC] using sources from the last 24 months only. Break this into market size and growth rate, the main players and how they position themselves, and the 3–5 trends that will matter most over the next 12–24 months. Avoid speculation and marketing language. Structure the answer with clear headings and end with a short “So what?” explaining why this matters in practice. Cite all claims. 2. Competitive Comparison Breakdown Compare [COMPANY A], [COMPANY B], and [COMPANY C] and explain how they actually compete. Start with a table comparing core offering, target customer, pricing model, distribution channels, strengths, and weaknesses. Then write a short summary explaining who looks best positioned and why. Call out where information is unclear or contradictory. Use verifiable sources only and cite everything. 3. Trend Validation Check Validate whether [TREND / CLAIM] is genuinely happening or being overstated. Gather evidence supporting it, evidence pushing against it, and real-world examples of adoption. Prioritise data over opinion and use recent sources only. Finish with a verdict: hype, early signal, or established shift, with a brief explanation. Cite all sources. 4. Deep Dive on a Single Question Research and answer this question in depth: [INSERT SPECIFIC QUESTION]. Pull from multiple independent sources and avoid surface-level summaries. Explain where experts agree, where they disagree, and what nuance or edge cases are often missed. Structure the answer clearly and be explicit about uncertainty. Cite all claims. 5. Buyer & User Insight Synthesis Analyse how real customers talk about [PRODUCT / CATEGORY] using reviews, forums, case studies, and public discussions. Ignore brand marketing pages. Summarise the most common pain points, desired outcomes, objections, and the language customers repeatedly use to describe the problem. Focus on patterns rather than anecdotes. Cite representative sources. 6. Regulation & Risk Overview Provide a clear overview of the regulatory, legal, and compliance considerations for [INDUSTRY / ACTIVITY]. Focus on what affects real decisions rather than theoretical risk. Include key regulations by region where relevant, current enforcement trends, upcoming changes to watch, and the practical implications for operators. Keep it factual and neutral. Cite all sources. 7. Evidence-Based Opinion Builder Help me form a defensible opinion on [TOPIC / POSITION]. Lay out the strongest arguments on both sides using data, expert commentary, and real-world outcomes. Go beyond surface-level takes and highlight where evidence is strong versus where it is mixed or inconclusive. Then summarise what the evidence supports today and where uncertainty remains. Cite everything. 8. Research-to-Decision Summary Based on current research, data, and expert commentary, summarise what someone should do next about [DECISION / TOPIC]. Structure the output as: what is known for sure, what is still unclear, key risks or trade-offs, and sensible next steps. Keep it practical and avoid predictions without evidence. Cite all sources and flag where judgement is required. ❤️ Like 🔁 Retweet 🔖 Bookmark Follow👤 @sakhil_ai for more such posts
@AravSrinivas ·
Claude Opus 5.5 is now available to all Perplexity Computer users. On our Wide-And-Deep-Research evals, it compares favorably to Fable 5.1 at a fraction of the cost. Opus 5.5 will become the “Standard” Effort orchestrator on Computer for all Pro and Max users. Congrats @AnthropicAI!
@xenovacom ·
NEW: Perplexity just released pplx-embed, a series of powerful embedding models. Inspired by their logo, I built a small semantic search demo: enter a query and the model retrieves the most relevant page. Everything runs locally in your browser with Transformers.js and Three.js
Watch video
@GergelyOrosz ·
The ongoing degradation of Perplexity needs to be studied. A year ago, this search engine did not hallucinate as much as today. I asked what World Cup game was on last Wednesday, and it hallucinated that there were none, after searching and finding England vs Argentina...
@VaibhavSisinty ·
NVIDIA is in talks to invest in Perplexity at a $30 billion valuation. Here's why. → Perplexity's revenue tripled in eight months from under $250M to over $750M → The growth is coming from Perplexity Computer an AI agent that does real work on your computer. Runs code. Uses tools. Completes tasks. → AI agents need GPUs for thinking and CPUs for doing. NVIDIA built its first CPU called Vera specifically for agents. Perplexity tested it. 1.5x faster. → Jensen first invested in Perplexity at $520M in 2024. Re-invested every round since. Uses it daily. → Before these talks, NVIDIA tried to pay billions to license Perplexity's tech. Settled on equity instead. $520M to $30B in under three years. NVIDIA isn't backing a search engine. They're backing their biggest proof that AI agents and their new CPU actually work.
@shushant_l ·
I'm amazed most people still use AI search like a basic chatbot. Here's how to use Perplexity like a professional researcher with one complete guide. --- 1. Perplexity combines powerful AI models with live web search and source citations for accurate answers. --- 2. Use it for research, learning, coding, writing, SEO, market analysis, and business decisions. --- 3. Explore features like AI Search, Deep Research, Labs, Pages, Spaces, file uploads, and image generation. --- 4. The free plan includes AI search, live web search, citations, follow up questions, and basic file uploads. --- 5. The Pro plan unlocks premium AI models, Deep Research, Labs, larger uploads, and higher usage limits. --- 6. Deep Research searches multiple sources, compares viewpoints, verifies facts, and creates detailed reports. --- 7. Labs helps build reports, presentations, dashboards, spreadsheets, charts, code, and interactive content. --- 8. Spaces keeps research organized into separate workspaces for teams, businesses, students, and long term projects. --- 9. Pages turns your research into polished webpages, reports, tutorials, documentation, and knowledge bases. --- 10. Upload PDFs, DOCX, PPT, CSV files, and images to summarize, compare, and analyze documents. --- 11. Use a structured prompt with Role, Goal, Context, Requirements, and Output Format for better results. --- 12. Create research prompts that compare viewpoints, include citations, summarize findings, and generate recommendations. --- 13. Generate blogs, LinkedIn posts, X threads, newsletters, scripts, and FAQs with content prompts. --- 14. Build SEO content by researching search intent, keywords, FAQs, titles, and meta descriptions. --- 15. Improve coding by explaining code, finding bugs, optimizing performance, and increasing readability. --- 16. Learn faster by asking for beginner to advanced lessons, examples, quizzes, and summaries. --- 17. Analyze businesses with SWOT analysis, competitors, risks, opportunities, and future outlooks. --- 18. Always be specific, provide context, ask for citations, upload documents, and verify important facts. --- 19. Avoid vague prompts, ignoring citations, skipping source verification, and mixing unrelated questions. --- 20. Follow the complete workflow from AI Search to Deep Research, Labs, Pages, Export, and Share for maximum productivity. --- To learn more, check the infographic. ---
@MollySOShea ·
NEW: I sat down w/ Perplexity Founder & CEO Aravind Srinivas (@AravSrinivas) on why AI-native SMBs could drive the next American millionaire economic boom. @perplexity_ai, valued at ~$20B with ~$2B in reported funding, $750M in annualized revenue as of August, seeing 5x YoY revenue growth as it expands from search into autonomous agents. One example Aravind shared is a single-person company generating $3M in annual revenue while spending over $1M a year on Perplexity Computer. His thesis is that AI could fundamentally change the economics of building a business. Smaller teams, less capital, higher productivity, & entirely new business models. At the infrastructure layer, Aravind sees a shift toward local & hybrid inference, with AI agents operating across personal devices and the cloud. As context windows expand and workloads become more personalized, local compute becomes increasingly cost-competitive, while offering greater privacy, control & data sovereignty. We cover: › Why SMBs could become a major driver of US GDP growth › The shift from AI models to agents & recursive economics › Why Perplexity is betting on your Mac instead of building data centers › NVIDIA, Apple Silicon, local inference & sovereign AI › Why writing code is no longer the bottleneck, but problem-solving is 𝐓𝐈𝐌𝐄𝐒𝐓𝐀𝐌𝐏𝐒 (00:00) Aravind Srinivas, Co-founder & CEO of Perplexity (01:30) Why Perplexity's brand is built on curiosity (04:47) Every great invention starts with one question (08:18) How to build a $1T company in America alone (12:36) The next big companies will be tiny (13:45) Is the American dream still alive? (16:28) Stop blaming users for bad prompts (22:55) How Perplexity quietly grew 5x (26:51) The solopreneur spending $1M a year on Perplexity (28:28) "We don't have competitors, we have contributors" (29:20) The data center is coming to your laptop (34:32) Orchestrating up with NVIDIA, down with Apple (36:11) What happens when your AI agent knows everything (39:17) Why Perplexity is skipping the data center race (41:25) Stop asking AI to write code, ask it to think (45:47) Perplexity’s answer to AI security (48:27) The Hugging Face incident & multi-agent threats (50:45) Turning everyone into a scientist (53:31) Why Perplexity isn’t chasing a billion users
Watch video@shushant_l ·
I'm amazed most people still use Perplexity like a basic search engine. Here's how to prompt Perplexity AI like a researcher to get dramatically better answers. --- 1. Perplexity is an AI answer engine that searches the live web and cites its sources. --- 2. Use it when you need current information backed by references. --- 3. Start every prompt with a clear action verb like summarize, compare, or analyze. --- 4. Always provide enough context instead of asking broad generic questions. --- 5. Add a specific timeframe to keep results relevant and up to date. --- 6. Tell Perplexity exactly how you want the output formatted. --- 7. Avoid vague prompts because they often produce vague answers. --- 8. Use Standard Search for quick factual lookups. --- 9. Use Pro Search for multi-step questions that need deeper reasoning. --- 10. Use Deep Research for reports, presentations, and comprehensive analysis. --- 11. Use Learn Mode when you're studying a completely new topic. --- 12. Use Model Council to compare answers from multiple AI models. --- 13. Choose the right source pool like Web, Academic, Reddit, YouTube, or SEC filings. --- 14. Add recency requirements whenever fresh information matters. --- 15. Ask Perplexity to challenge assumptions and present counterarguments. --- 16. Specify trusted source types such as peer-reviewed papers or official websites. --- 17. Request source verification for every major claim it makes. --- 18. Continue the conversation with follow-up questions to refine the answer. --- 19. Always open and verify important citations before trusting them. --- 20. Think like a researcher by defining the scope, evidence, and output before prompting. --- To learn more, check the infographic. ---
@Jack_Raines ·
Perplexity Computer lowkey cooks. I've dunked on them a lot for the "let's buy Chrome" and other social media shenanigans, but this thing rips. Been working on overhauling a personal website and needed to do a really tedious re-labeling of blog titles/dates. Claude could do it, Perplexity one-shotted it.
@AiTechPerks ·
Just built my 2026 personal investment research assistant in Perplexity Computer. Every Monday it now automatically: • scans latest filings & earnings for my full 25-stock watchlist • runs sentiment analysis across X, Reddit & news • checks insider trading + short interest changes • flags any position >10% off the 50-day MA • then drops a clean 1-page PDF summary with buy/sell/hold signals + confidence scores What used to take me half a day is now done before my first coffee. First run already looks sharp.
Watch video
@Suryanshti777 ·
🚨 Perplexity just quietly launched something big. Not a chatbot. Not an agent. An AI computer. This changes how AI gets used every day. Here’s what “Perplexity Computer” actually does: • Routes tasks across multiple models automatically • Runs sub-agents in parallel • Executes code in a sandboxed environment • Connects to Gmail, Slack, Notion, Drive, GitHub • Remembers context across sessions • Schedules recurring autonomous tasks • Generates docs, dashboards, apps, and reports • Produces images + video inside workflows You don’t pick models anymore. You describe outcomes. It builds the workflow itself. Example prompts: → “Build a competitor dashboard for my SaaS” → “Research top 5 CRM platforms and compare pricing” → “Generate my weekly content calendar from trends” → “Track my watchlist and email me daily summary” → “Create a 4,000-word SEO page + publish” Behind the scenes: 1. Breaks goal into task graph 2. Assigns best model per task 3. Spawns sub-agents in parallel 4. Runs tools + code execution 5. Self-heals on errors 6. Ships final deliverables This is the important shift: Old AI → answers questions New AI → runs workflows Old AI → one prompt New AI → persistent system Old AI → chat interface New AI → operating layer We’re moving from: “Ask AI” to “Assign AI” Perplexity Computer is basically: Your AI analyst Your AI researcher Your AI developer Your AI automation engine Running 24/7. The interface didn’t change. The architecture did. And most people haven’t realized it yet.
@thecolbykultgen ·
7 manual tasks Perplexity's AI browser now handles for me: 1. Unsubscribe me from all marketing emails 2. Compare products across multiple tabs to find the best deals 3. Accept LinkedIn connection requests that have more than 10 mutual connections 4. Find important emails I forgot to reply to and flag them 5. Plan trips and create detailed travel itineraries on Google Maps 6. Summarize YouTube videos and pull any valuable links or resources mentioned 7. Search Reddit, YouTube, and Google simultaneously for the most reliable answer Yes, it can do all of that (and more). Comet is @perplexity_ai 's AI-powered browser that handles everything from email to research to shopping. You can give it instructions, and it autonomously completes multi-step tasks in the background while you focus on other work. It's going to be a game-changer. Try it out here: https://t.co/Q853ENOI72
Watch video
@harshitagu72595 ·
Perplexity Computer just did the kind of competitor research people waste half a day on. Pricing positioning messaging market signals All pulled into one usable output. This is where @perplexity_ai starts to feel genuinely useful. https://t.co/SSQ5kFR07o
Watch video
@techxutkarsh ·
Built and deployed a production-ready web app using Perplexity Computer with authentication, database, and a clean UI. What stood out wasn’t just speed, but how clearly the full system comes together: - Frontend for user experience - Backend for logic and APIs - Database for storage - Auth layer for security As it scales: - Traffic gets distributed - Services grow independently - Performance improves with caching If you’re exploring how real-world apps are built and scaled, you should try Perplexity Computer.
Watch video
@thisdudelikesAI ·
A kid from Chennai built the company that Google's CEO declared a Code Red emergency over, and he is 32 years old and still codes daily. His name is Aravind Srinivas. He runs Perplexity. Here is the story, because the version most people know is wrong on almost every detail. Aravind grew up in a vegetarian household in Chennai, the same city Sundar Pichai grew up in. He trained for the IIT entrance exam with his friends. They all got into computer science at IIT Madras. He missed the cutoff by a fraction of a percentage point and got placed in electrical engineering instead. He says he was depressed for a year and mostly hung out with his friends who got the seat he wanted. He stayed at IIT Madras and finished a dual BTech and MTech in electrical engineering. Then he flew to Berkeley for a PhD in computer science. He published heavily on reinforcement learning, contrastive learning, and transformers. He worked on the Decision Transformer paper. He built VideoGPT. He interned at DeepMind in London. While he was there he read a book called "In the Plex" about Google's first 15 years. The book changed something. He decided he wanted to build a company that could one day stand next to the one Pichai built. He went to OpenAI as a research scientist. He contributed to DALL-E 2. He worked at Google. Then in August 2022, four months before ChatGPT launched, he co-founded Perplexity in a small apartment with Denis Yarats from Meta FAIR and Andy Konwinski, a co-founder of Databricks. He built the first prototype over a single weekend. The idea was simple. Search engines force you to read 10 blue links and figure out the answer yourself. Aravind hated that. He wanted a search engine that gave you the actual answer with citations underneath, like a research assistant that did the reading for you. Jeff Bezos invested in the seed round. Nat Friedman invested. Susan Wojcicki invested. They had $3 million and a small team. In December 2022 ChatGPT launched. Three weeks later Sundar Pichai declared a Code Red inside Google. Internal memos called AI search an existential threat to the company. Pichai personally called founders back to work on Google's response. The man Aravind had grown up idolizing was now panicking about the exact category Aravind had just started a company in. Perplexity grew faster than anyone expected. The product worked. Researchers used it. Investors used it. Journalists used it. By November 2024 the company was valued at $9 billion. By March 2025 it was raising at $18 billion. SoftBank led the round. Aravind became famous for not behaving like other CEOs. He answered DMs on X personally. He posted product ideas at 2 in the morning. He pushed code into production. He fought publicly with Google and OpenAI and called Sam Altman a competitor on stage. He went on every podcast that asked. Perplexity has now shipped its own browser called Comet. It has shipped an agent called Computer that can take actions across the web for you. It has signed enterprise deals with Snowflake and runs in dozens of Fortune 500 companies. India is now its largest user base. The kid who could not get into computer science at the college he wanted, who spent a year of his life depressed about it, built a company that now competes head-on with the search giant founded by the most famous engineer from his hometown. Aravind still works seven days a week. He still posts code commits. He still answers his own customer support tickets sometimes. He runs a company worth more than the entire market cap of most public tech firms, from an office where he claims he is in by 9 in the morning every single day. A failed entrance exam in Chennai became the most credible threat to Google Search in 25 years.
@TopherNOW ·
Finding tenants, planning tours, and pulling demographics data used to mean 3 different paid tools and half a day of manual work. @TristenPalori showed CRE AI Studio how he does all of it in minutes with @perplexity_ai Computer. Live with a couple of keystrokes, Tristen was able to 👇 🔍 Pull Census demographics for any site 📊 Find 20K SF office tenants with contact data 🗺️ Build a tenant tour plan with the Google Maps API 📄 Export client-ready site comparison PDFs on the spot The wildest part? He built a full tenant rep software from scratch using just Perplexity Computer. Tour routing, demographics, site comparisons, all in one tool. Robert Beringhaus from Perplexity's enterprise team also joined and broke down how the tech behind the latest features actually works. Huge thanks to @TristenPalori for breaking down what this thing can do. Check out the full session at https://t.co/fJ9iRAC95l (7 day free trials still available)
Watch video
@itsolelehmann ·
your audience already wrote your best marketing material, you just haven’t collected it yet i built a "voice-of-customer" workflow that makes all of my writing sound EXACTLY like my audience's thoughts which means more engagement, more followers, more sales… because people buy from people who understand them hormozi says it all the time: "he who is closest to the customer wins." this is how you get close. you stop guessing at what your audience is thinking and feeling and start reflecting their exact words and psychology back to them. it's one prompt in perplexity computer. gives you 50+ quotes from real people, organized by pain point theme so Claude can draw upon their exact phrases every time it writes for you here's how to set it up: 1. open perplexity computer (i use this because it's a search task at scale. perplexity computer is built on search infrastructure. we need to crawl across reddit, g2, trustpilot, product hunt, and indie hackers at once and pull exact quotes from hundreds of real threads, this is what it was literally designed for) 2. paste this prompt (swap in your own audience and topic): —— "I need you to build a voice-of-customer quote bank with 50+ real quotes from real people, organized by pain point theme. The pain points I'm looking for: what [YOUR AUDIENCE, ex: solopreneurs, founders, online business owners] are struggling with, frustrated by, excited about, or wishing for when it comes to [YOUR TOPIC, ex: using AI to run and market their businesses]. Search across Reddit, Product Hunt, G2, Capterra, Trustpilot, Indie Hackers, and any relevant communities. Read full reviews, threads, and comments, not summaries. I need the actual words people use, not your summary of what they said. Build the quote bank organized by pain point theme. Each theme should have a short label (2-5 words) and 15-25 quotes under it. Just the quotes, nothing else. Prioritize quotes that are specific, emotional, vivid, and sound like a real person talking. Skip anything that reads like a formal review or press release. I want the raw, unfiltered way people actually describe these problems. Include a separate section for breakthrough moments and wins. Quotes where people describe something that actually worked, a genuine shift, or a result that surprised them. And a final section that's just recurring phrases and slang this audience uses to describe their problems, grouped by theme. No full quotes needed, just the expressions themselves. Find at least 10 distinct pain point themes. Go deep rather than broad. I'd rather have 8 great quotes per theme than 2 quotes across 30 themes. Only include real quotes you can verify. Don't fabricate or paraphrase." —— 3. let perplexity cook. takes about 10 minutes 4. save the output as a context file in whatever ai tool you use most (perplexity, claude, chatgpt, whatever) 5. add one instruction to your writing setup: "reference the voice-of-customer file and use the audience's exact language for hooks, pain points, and framing" now every time you write, your ai pulls from how your audience actually talks instead of how you think they talk (honestly it feels like cheating) run it again every quarter to keep it fresh
@Freyabuilds ·
Competitor research always sounds easy until you’re 20 minutes in, drowning in tabs and screenshots. I tried running the whole thing inside Perplexity Computer instead. In this video I’m pulling a bunch of competitor pages into one workspace, asking it to surface the positioning, and then cleaning up the overlaps so I can actually see who’s saying what. It still takes some thinking on my side, but having the sources, summaries and my own notes in one place makes reviewing everything way less painful.
Watch video
@AITechEchoes ·
Been testing Perplexity Computer for a few days now and something clicked. Gave it a multi-step AI research task that normally eats up my entire morning - jumping between ChatGPT for drafts, Perplexity for sources, Notion for organizing, and then manually formatting everything into something presentable. This time I just described what I needed in one prompt. It broke the task into steps on its own, pulled real sources, organized the findings, and gave me a clean final output I could actually use no copy-pasting between tabs, no reformatting, no babysitting the process. Honestly didn't expect it to connect the steps on its own. I just described the end result and it figured out the order. That's the part I can't stop thinking about.
Watch video@perplexitydevs ·
A new Perplexity Search SDK cookbook is live. The recipe fans out focused searches, filters results to official docs, extracts relevant passages, and writes a source-linked brief for your coding agent. Get started: https://t.co/lJEWmx0YB1
Watch video
@TechByMarkandey ·
I just experimented with Perplexity Computer and it blew my mind. With a single prompt, I was able to generate a complete business plan including: - Executive Summary - Market Analysis - Go-to-Market Strategy - Financial Projections - Team Overview And the best part? It was instantly exportable as a shareable PDF. The potential here is huge. Tools like this aren’t just speeding up work they’re transforming how we create, plan, and execute ideas. It makes me wonder: what will the future of business planning look like when AI can handle the heavy lifting in minutes?
Watch video
@AIwithArsalan ·
I tried Perplexity Computer on a messy TXT file today and this was one of the more practical use cases so far. Instead of leaving me with a wall of raw text, it helped turn it into a structured DOCX report with a clear title, cleaner structure, and more organized sections. 1. Upload a raw file (CSV, PDF, DOCX, TXT) 2. Give it a simple prompt 3. And get an organised docx. file Perfect for reports, pitch decks, and client presentations. That’s the part I liked. It felt less like AI just rewriting text and more like turning rough work into something actually presentable.
Watch video@techxutkarsh ·
I just tried something cool with Perplexity Computer. I asked it to research my top competitors looking at their pricing, positioning, and where I could stand out and the insights were eye-opening. It’s amazing how AI can turn what used to take days into minutes. I now clearly see: - How competitors structure their pricing and why - How they talk to their customers and position themselves - Where customers feel underserved - Opportunities where I can differentiate and add real value Competitive research isn’t just about knowing who’s out there it’s about understanding the gaps you can fill. Using AI to organize all this information made the process so much faster and clearer. Curious how do you usually research your competitors?
Watch video
@JustAnotherPM ·
Perplexity is worth $22.6 billion. It doesn't even own a single AI model. Here's what makes it so powerful When you type a question into Perplexity, five things happen in about a second: 1. It parses your intent to figure out what you're actually asking. 2. It does live web searches using both keyword and semantic search. 3. It reranks the results by relevance. 4. It assembles the most relevant information into a prompt with citations. 5. It gives that prompt to an LLM, which is bound to answer by using the evidence it was given. But here's the part most people miss. 𝗣𝗲𝗿𝗽𝗹𝗲𝘅𝗶𝘁𝘆 𝗱𝗼𝗲𝘀𝗻'𝘁 𝗽𝗶𝗰𝗸 𝗼𝗻𝗲 𝗺𝗼𝗱𝗲𝗹. It routes your question to whichever model fits best. A "model-agnostic router" trained with reinforcement learning looks at speed, cost, and difficulty, then decides whether your question goes to GPT, Claude, Gemini, Grok, or their own in-house Sonar models built on Llama. Easy questions get the cheap fast model. Hard questions get the frontier model. Their newest feature, Model Council, takes it further. Multiple models answer the same question, and a synthesiser reconciles them into one response. This is the strategy that matters: Perplexity owns neither the models nor the content. Every time OpenAI or Anthropic or Google ships a faster, cheaper, smarter model, Perplexity's product improves for free. They are arbitraging a war between giants. The lesson I keep coming back to: you don't have to own the hardest layer. You have to own the layer that gets more valuable as the rest of the stack commoditises. Which layer are you building on? I wrote a detailed piece on Perplexity (linked below)
@aakashgupta ·
For $20/month and zero setup, you can now run parallel AI agents that deliver finished work while you sleep. Perplexity shipped Computer. Back on Ramp's fastest-growing B2B software list. 19+ AI models. 400+ connectors. The reason isn't search anymore. Every take I've seen focuses on the "AI assistant" framing. They're all underselling it. Computer doesn't give you suggestions. It delivers the finished thing. Research reports with source citations. Deployed dashboards with shareable links. Cleaned datasets with charts. Launch kits with positioning docs and email drafts. Three things make it different from everything else out there. Cloud execution, so your laptop can be closed. Parallel agents, so five tasks run simultaneously. And persistent memory, so you stop re-explaining yourself every session. I pointed it at Notion's product pages. 28 pages scored across 5 criteria, competitive benchmarks against Coda and Slite, with specific recommendations per page. That's a $15K messaging audit. Took about 20 minutes. But credits disappear fast if you don't know how to prompt it. I burned hundreds learning this. Built a five-rule Prompt Spec that cuts cost by 60%+. I spent weeks testing it. Today's guide has the six PM use cases, exact prompts, the credit-saving system, and an honest comparison against Claude Code, Cowork, and OpenClaw. Full guide: https://t.co/xHaRK91SEA
@JustinThomasAI ·
I have no idea why more people do not talk about Perplexity. It is by far my favorite daily AI. One cool feature of many: When you have the app on your phone, you get news article notifications for topics you actually care about straight to your lock screen. Just clean, sourced news delivered like a smart assistant who knows what you want to read. The Discover section inside the app is even better. Hand-picked stories based on your history, AI-summarized, with full sources. No other AI has built anything close to this.
@AIwithArsalan ·
I tried Perplexity Computer on a full landing page workflow today, and this is where it started clicking for me. I wasn’t just using it for a headline or a few quick lines. I pushed it to help shape the whole page, different headline directions, hero copy, benefits, features, FAQs, CTA variations, even A/B test ideas. That’s what made it feel useful. It didn’t feel like random AI copy pasted onto a page. It felt more like getting to the full message and structure much faster. Most people are still using AI to chat. Perplexity Computer gets a lot more interesting when you use it to actually do the work.
Watch video
@HeyZaraKhan ·
🚨BREAKING: Perplexity AI just turned into a full personal finance OS. Users can now connect Perplexity Computer with Plaid (the world’s largest financial data network). This is like a complete financial picture in one place. Here’s what this means for everyone: → Full financial visibility: Connect bank accounts, credit cards, loans, and investments in one place. → Real-time net worth tracking: Auto-calculates and updates across all linked accounts → AI-powered spending analysis: Tracks every transaction and categorizes spending instantly → Custom dashboards: Ask it “Build my monthly budget” → it creates it live → Debt payoff planner: Analyzes interest rates + balances → gives a payoff strategy → Retirement intelligence: Tracks your progress based on income, age, and investments → Cash flow forecasting: Predicts low-balance weeks before they happen → Ask anything about your money: No dashboards needed, just ask questions in plain English → Institution-grade data: Powered by sources like Nasdaq, S&P Global, and U.S. Securities and Exchange Commission → Secure by design: Read-only access. Your financial data never touches Perplexity servers I hope you found this informative and helpful for more you can follow me @HeyZaraKhan
@Suryanshti777 ·
🚨Breaking: Perplexity AI just introduced Perplexity Computer — and it flips AI from answers → execution. This isn’t a chatbot. It’s a job runner. Developers already had this with autonomous agents. Now non-developers can do it from a browser. The shift is simple but massive: Wrong way: “What are risks in this contract?” New way: “Review this contract, fact-check claims, flag risks, add tracked changes, export Word doc.” Same input. Completely different outcome. People are already using it to: • clean messy CSVs automatically • build dashboards from raw data • audit competitor content • research leads and send emails • review legal contracts • generate weekly reports on schedule One user automated a Monday task: 90 minutes of spreadsheet cleaning → 4 minutes Then scheduled it to run every week. Without opening their laptop. Another example: Find companies advertising on competitor podcasts → identify decision makers → write personalized emails → send them automatically No prompting. No babysitting. Just delegation. This changes the skill entirely: Old skill → writing prompts New skill → describing outputs The people who learn this early will have a massive advantage. Because AI isn’t just assisting anymore. It’s doing the work.
@emilylai ·
Testing Perplexity Computer and impressed. It has nine subagents simultaneously running deep research and analysis for me Cost $16 (1600 tokens) to produce a report on if there are any trade setups I should take - and if so, why, thesis invalidation, r:r, setup
@itsolelehmann ·
perplexity understood that by adding any “add your api” steps you lose 98% of potential users non technical users just want to use stuff this is why perplexity computer is dope imo yes you can build it in openclaw but most people just don’t want to
@teja2495 ·
After I saw the news about Mac Pro being discontinued on X, I immediately searched for the news on ChatGPT, Perplexity, Claude and Gemini to see which one has the latest news. Claude and Gemini were able to get the latest news about it being dead, but ChatGPT and Perplexity did not know about this news yet. I'm actually surprised that Perplexity is slower in fetching latest news.
@JustinThomasAI ·
My favorite thing about Perplexity: Comet browser. The in‑context chat panel that is docked alongside your active tab and can see that tab’s content is a game changer. It functions as a side panel assistant that can “co‑browse” or work in the context of whatever you’re reading.
@DivyanshT91162 ·
Most AI agents still search the web like this: Search → Read → Search Again → Read Again Perplexity just said: "why not write code instead?" Their new Search as Code architecture generates Python that directly calls the search stack, reducing the need for endless tool-calling loops. Now available in the Perplexity Agent API and already powering Computer by default. This feels like another step toward agents that execute workflows, not just chat. Full breakdown 👇
@TheGeorgePu ·
Can really feel Perplexity in revenue push mode now. - Pausing daily tasks after light usage. - Defaulting every query to Deep Research. - Pushing Perplexity Computer hard. Great product. Ugly math.
@dino11 ·
Perplexity launched "Model Council" — a feature that runs your query through GPT-5.2 and Claude 4.6 simultaneously and lets you compare the outputs side by side. This is how you should be evaluating AI tools for your business: Never trust one model's answer. Run the same prompt through 2-3 models. Compare. The differences will shock you. Especially for customer-facing content where accuracy matters. Free on Perplexity Pro.
@glenngabe ·
Based on AI agents cloaked as human users, which threatens ad revenue -> US publishers back Amazon in AI agent access dispute with Perplexity "Perplexity is now appealing that decision. It has argued that its AI agents are “much more transparent and limited” than Amazon’s own use of agentic AI to complete transactions on third-party retailer sites." "Amazon alleges that Perplexity has “purposely configured” Comet AI so that when the AI tool is deployed on behalf of an Amazon customer, “Perplexity falsely identifies its Comet AI agent activity as coming from Google Chrome, which is a separate, widely used web browser owned by Google. As a result, Perplexity’s Comet AI agent covertly poses as a human customer shopping in the Amazon Store on a Google Chrome browser.” “If publishers cannot distinguish between an AI agent and a human subscriber, they will not be able to measure audiences accurately.” https://t.co/5zFEGgKBcM
@Tegadesigns ·
🔧 Breaking: Perplexity Computer Just Got Bank & Investment Access A new feature just launched. Here is everything you need to know about it as a founder, creator, and professional : 🔹 What it’s used for → Perplexity’s Computer (the smart AI agent) now safely connects to your real bank accounts, credit cards, loans, and investments using Plaid. Plaid is the same secure system used by many popular finance apps, Venmo, Cashapp, etc. Once linked, you talk to it in normal English. Examples: “Show me my spending on eating out last month,” “Build me a plan to pay off my debt faster,” or “What is my full net worth right now?” It pulls all your data together and creates custom charts, budgets, debt plans, or retirement forecasts using your actual numbers. 🔹 How to implement it → It is live now on desktop for signed-in users in the US and Canada. Any signed-in user can link accounts safely with Plaid. Just sign in, connect your accounts, and start asking questions. Basic linking works for everyone. 🔹 Note : It is read-only, it can look at your data but never spend or move money. Your data never gets stored on Perplexity’s main servers. 🔹 Risks to know → - You are giving an AI access to very personal financial data, even if it is read-only and uses secure Plaid connections. - Always review what you share and consider starting with only one or two accounts to test. Never treat AI advice as professional financial advice, double-check important decisions with a real advisor. Follow and put on notifications 🔔 if you want to stay ahead on the latest tech and AI tools breakdowns.
@JulianGoldieSEO ·
Perplexity just killed the "one-person AI" era. This update quietly turns AI from a chatbot into your team's operating system. What's New: → Projects gives your entire team one shared workspace. → Files, memory, and AI sessions now live in one hub. → Everyone works from the same context instead of starting from scratch. Why It Matters: ✓ Built on Perplexity Brain's living context graph. ✓ Perplexity says repeat tasks became 25% more accurate, wasted searching dropped 16%, and repeat task costs fell 13%. ✓ Connect Slack, Salesforce, HubSpot, and Snowflake once, then everyone in the hub can use the same connections. The Catch: ✔ Shared memory also means shared mistakes. ✔ Someone should review and correct team memory early so the system improves over time. The next generation of AI isn't one smarter chatbot. It's one shared brain for your entire business.
Watch video@JulianGoldieSEO ·
Want to use Perplexity Computer with Snowflake? Start here: 1. Open connector settings. 2. Add the Snowflake integration. 3. Set up service account access. 4. Use key pair authentication. 5. Trigger data map generation. 6. Let it learn your schema. 7. Ask business questions in plain English. That’s how you turn Snowflake into an AI-powered analyst. Save this video, you’ll know how to connect Perplexity to your data.
Watch videoBest Perplexity tweets
Xholic studies what works in your niche, drafts posts in your voice and schedules them for the hours your audience is online.
No card needed · Nothing to cancel
Browse all tweet collectionsLive demo: no sign-up or login needed · takes about 2 minutes
Keep exploring