# Elephas

Aiolum story: https://aiolum.com/story/elephas/ · added 2026-10-26 · updated 2026-10-26

**In short.** Elephas (Mac app that answers from your own documents, India): US$150K a year, January 2026. Evidence grade D: reported, not verified. Team: 2 co-founders, Kamban S (builds the app) and Ayush Chaturvedi (marketing); the company blog also lists a content marketer. Launched 9 June 2022; still running.

Mac app that answers from your own documents · India · AI-native · team of 2 to 10 · still running · group: Documents, data and slides

- Outcome: US$150K a year, January 2026 (evidence grade of the headline number: D; headline in US dollars: 150000)
- Launch: 9 June 2022; months from launch to the outcome: 43
- Team: 2 co-founders, Kamban S (builds the app) and Ayush Chaturvedi (marketing); the company blog also lists a content marketer
- Outside funding: None found
- Main factor: Search traffic
- Meters, 1 to 4: evidence strength 1, copyable pattern 2, capital needed 1
- Tags: AI-native, Small team, Bootstrapped, Running, B2C, Productivity

**The product.** A desktop app for the Mac that reads the user's own files (PDFs, Word documents, spreadsheets, notes, web pages) and answers questions about them, with a citation to the page each answer came from. The files stay on the computer. The user chooses which AI writes the answer: the company's own service, a model provider paid with the user's own key, or a model that runs on the Mac.

**Why it worked.** Ayush Chaturvedi said in a Starter Story interview published in January 2026 that the app made US$150,000 a year, and that growth came in two phases: posts in Reddit communities first, then pages that rank in Google. The first version came out in June 2022, months before ChatGPT, as a writing helper that worked inside every Mac app. When chat apps became free, the founders moved the product toward answering from the user's own documents, a job the general chat apps did not do on the computer.

Evidence grades: A = disclosed by a public company, a filing or a payment processor, or verified by one; B = two or more independent reports agree, or a journalist confirmed it; C = a founder claim that no one else verified; D = reported, not verified: one founder claim or a third-party estimate with no second source; calc = computed on this page from graded facts.

## Summary

Elephas is a Mac app that turns a person's own files into a knowledge base, a searchable collection of documents that an AI can read before it answers. The app calls one collection a "Brain". You add PDFs, Word files, notes or web pages, ask a question in plain language, and the answer cites the file and page it came from. The documents stay on your computer. Only the question and a few matching passages go to an AI model, after names, email addresses and phone numbers are replaced by placeholders. You can also use your own API key, a password that lets a program use a model provider such as OpenAI on your own account, or a local model, an AI model downloaded to the Mac, so that nothing leaves the computer.

Kamban S, a developer in India, posted the first beta on Hacker News on 9 June 2022 as a writing assistant that worked inside Mail, Pages, Notes and other apps, with each user's own OpenAI key. In November 2022 he told Analytics India Magazine that it had reached US$1,000 in monthly recurring revenue. Ayush Chaturvedi joined as a marketing co-founder. In a Starter Story video published on 21 January 2026, he said the app was making US$150,000 a year, and the video title put it at US$13,000 a month. The app is sold by Liberation Softwares LLP, a company registered in Chennai, India.

The US$150,000 figure is a founder statement in one interview, so this page grades it D, reported and not verified. No payment processor, buyer or journalist has confirmed it. In 2026 the company replaced the app that earned that money with a new desktop app that only answers from documents and is sold by monthly plans. The page also shows what happened to earlier companies that sold a knowledge or writing tool next to products that large platforms gave away.

## Pricing

Elephas prices in October 2026, from the company's pricing page and the App Store listing of the earlier app (grade C)

| Plan | Per month, billed yearly (US$) | Per month, billed monthly (US$) | Limit |
|---|---|---|---|
| Personal | 15 | 19 | 1 computer; Elephas AI included |
| Professional | 24 | 29 | 3 computers; twice the included AI usage; encrypted sync |
| Team | 24 per person | 29 per person | shared Team Brains; the owner counts as one seat |
| Earlier app, Standard (App Store) | 15.83 | 19 | billed US$190 a year |
| Earlier app, Pro (App Store) | 32.50 | 39 | billed US$390 a year |
| Earlier app, Pro Plus (App Store) | 40.83 | 49 | billed US$490 a year |

The new desktop app has a 7-day free trial with no card. Personal and Professional include a monthly amount of Elephas AI; after that the user can continue with an own API key or the on-computer model. Team plans use Elephas AI only. The earlier app is no longer sold on the company's website but is still sold on the App Store at the prices below. In June 2025 the website sold the earlier app from US$8.99 a month, with a lifetime licence at US$129.

## How much revenue does Elephas make? The numbers, graded

A Mac writing and document assistant that the co-founder said made US$150,000 a year in January 2026, with one earlier figure from November 2022

| Fact | Value | Grade | Note | Sources |
|---|---|---|---|---|
| Revenue a year, as stated in January 2026 | US$150,000 | D | Ayush Chaturvedi in the Starter Story video published on 21 January 2026; the video description reads "Ayush built Elephas, an AI app that's making $150K/year". One interview, no second source. The recording date is not stated. | [1, 2] |
| Revenue a month, video title | about US$13,000 | D | Video title "How I Built a $13K/Month SaaS". ProvenStartups, which watched the video, writes "about $12K/month currently". | [1, 2] |
| Monthly figure from the yearly one | US$12,500 | calc | Calculated: US$150,000 divided by 12. It agrees with the US$12,000 to US$13,000 a month said in the same video. | [1, 2] |
| Revenue screens shown on camera | about US$110,000 on Gumroad for the current year and about US$12,000 on the Mac App Store | D | As described by ProvenStartups from the video. A founder showing his own dashboard; this page could not view the screens itself, and the period of "current year" is not stated. | [2] |
| Revenue attributed to Google search, 2023 | about US$70,000 | D | Co-founder in the video, as summarised by ProvenStartups. | [2] |
| Monthly recurring revenue after six months | US$3,000 | D | Co-founder in the video, as summarised by ProvenStartups: Reddit posts brought the first customers. | [2] |
| Monthly recurring revenue, November 2022 | US$1,000, with more than 200 people using it every day | C | Kamban S to Analytics India Magazine, 15 November 2022, about five months after the beta. The article says Kamban is the name he goes by. | [3] |
| Website visits | about 180,000 a year | D | Cited in the video, as summarised by ProvenStartups. | [2] |
| People using it | more than 3,000 | C | "3000+ people trust Elephas" on the June 2025 pricing page; "Trusted by 3,000+ professionals" on the October 2026 home page. | [12, 7] |
| First public version | Beta posted on Hacker News, 9 June 2022, 21 points | C | Kamban S: "Yes, users need to input their own OpenAI API keys." | [6] |
| Prices in June 2025 | from US$8.99 a month; lifetime licence US$129 | C | Wayback copy of the pricing page, 24 June 2025. Pro was US$14.99 and Pro+ US$18.99 a month. | [12] |
| Prices in October 2026 | US$19 a month (Personal) to US$29 per person a month (Team) | C | Company pricing page. About 20% less when billed yearly. | [8] |
| Where the earlier app was sold | own website (licences through Gumroad), Setapp, Mac App Store, iOS App Store | C | Terms of use name Gumroad and the Apple App Store; ProvenStartups lists Setapp from the video. | [10, 2] |
| Company | Liberation Softwares LLP, India, founded 2021 | C | About page, terms of use and privacy policy. The terms are governed by the laws of India. | [9, 10] |
| Team | 2 co-founders; the blog lists a content marketer | C | Ness Labs names both co-founders. Kamban S told Website Planet: "We don't want to have more than 10 people." No source gives a headcount at January 2026. | [4, 5, 15] |
| Outside funding | none found | C | No source read for this page mentions investors or a funding round. | [2, 4, 5] |
| Latka's revenue estimate, June 2024 | US$15 million a year | D | A third-party estimate; Latka says the company "has not published its revenue". It is 100 times the founder's own figure and should be ignored. | [16] |
| Time from the beta to the January 2026 figure | 43 months | calc | Calculated: June 2022 to January 2026. | [6, 1] |

How to read this table: grade D values rest on one founder interview, or on a third-party estimate with no second source. Grade C values are statements by the founders or the company's own pages. A "calc" value was computed on this page from the rows above it. No value is grade A or B.

## When did Elephas launch, and how did it grow? Timeline

From a beta posted on Hacker News to a reported US$150,000 a year, in 43 months. Each entry carries its grade.

- **9 June 2022** (grade C [6]): Kamban S posts "Show HN: Personal AI Writing Assistant for Mac". It fixes grammar, rewrites sentences and drafts emails inside other Mac apps, using each user's own OpenAI key.
- **July 2022** (grade C [6]): A second Show HN post for "Elephas V2", and a third that links to the app's Gumroad page.
- **15 November 2022** (grade C [3]): Analytics India Magazine reports US$1,000 in monthly recurring revenue and more than 200 daily users. The app is written in Swift and does not send text to the company's servers.
- **2023** (grade D [14, 2]): Ayush Chaturvedi is co-founder for marketing; Starter Story's summary dates his start to 2023, while he told Ness Labs that Kamban asked him the day after the first launch. In the 2026 video he said that Reddit brought the app to US$3,000 a month in six months and that Google search brought about US$70,000 in 2023.
- **7 February 2024** (grade C [4]): Ayush Chaturvedi tells Ness Labs that "Super Brain", chat with a collection of one's own files and notes, is the most loved feature.
- **24 June 2025** (grade C [12]): The pricing page sells plans from US$8.99 a month and a lifetime licence at US$129, and says 3,000+ people trust Elephas.
- **21 January 2026** (grade D [1, 2]): Starter Story publishes the interview in which Ayush Chaturvedi says the app makes US$150,000 a year, about US$12,000 to US$13,000 a month.
- **October 2026** (grade C [7, 8, 10, 11]): A new desktop app answers only from documents, adds a Team plan and its own Elephas AI service on open-source models. The earlier app is no longer sold on the website.

## Why did Elephas work?

### Inside every app, not in a browser tab

The first version sat in the Mac menu bar and worked on selected text in Mail, Pages, Slack or Google Docs. In June 2022 the large models were only reachable through an API, so a native app with a keyboard shortcut was a real improvement over copying text into a web page.

### The user's own key

From the first day, users entered their own OpenAI API key. The company did not pay for model usage and did not store users' text, so a small team could sell it at a low price and keep most of the revenue.

### Reddit first, Google search later

In the 2026 interview the co-founder said that short demonstrations in niche Reddit communities brought the first US$3,000 a month, and that search pages later grew revenue to about US$12,000 a month, as ProvenStartups reported. The company blog now has close to 300 posts, many of them comparisons and lists of AI tools.

### Answers from your own files

When ChatGPT became free, a writing helper alone was easy to replace. The founders put the weight on "Super Brain": answers from the user's own documents, kept on the computer, with citations. The 2026 app does only this.

## How it was built

These details come from the founders' interviews and the company's own pages. They are grade C.

- A native Mac app written in Swift, according to Kamban S in November 2022. Later versions ran on the iPhone and iPad, with an AI keyboard for iOS.
- The earlier app sent requests to the model provider that the user chose, with the user's own key. The privacy policy lists OpenAI, Anthropic and Google for that app, plus local models for offline work.
- The 2026 app has three routes: Elephas AI, which the privacy policy says runs on open-source models in the company's own Amazon Web Services environment behind a gateway on Cloudflare; the user's own key for OpenAI, Claude, Gemini, Groq, Mistral and others; or a downloaded model of about 1.3 to 5.7 GB that runs on the Mac.
- Before a question leaves the computer, names, email addresses and phone numbers are replaced by placeholders and restored in the answer. The company says this works best in English and is not a guarantee.
- Account records are kept by Supabase and anonymous usage reports by PostHog, according to the privacy policy.
- Licences for the earlier app were sold through Gumroad, the Apple App Store and Setapp. Kamban S told Analytics India Magazine that he made the product demo videos with AI.

Sources: [3, 10, 11, 13, 2]

## Who tried this before Elephas, and why did they fail?

Five companies in the Loot Drop database sold a knowledge, writing or search tool that a large platform or a free AI product later made unnecessary. The values below are from that database and are not independently verified.

The failed startups come from Loot Drop (https://www.loot-drop.io/), a free database of dead startups, retold here in our own words; Loot Drop describes its entries as AI-assisted summaries of public sources, so they carry no evidence grade.

### Draft (2011 to 2014; funding US$120K; cause: Competition)

- Promise: A simple writing and editing tool for writers.
- Why it died: Google Docs added the same features for free inside a suite that people already used.
- Record: https://www.loot-drop.io/startup/1605-draft

### Kippt (2012 to 2017; funding US$400K; cause: Competition)

- Promise: Save and organise links to fight information overload.
- Why it died: Pocket and Evernote covered the same need with more features, and its social sharing did not bring new users.
- Record: https://www.loot-drop.io/startup/278-kippt

### Niles (2017 to 2020; funding US$2.8M; cause: Competition)

- Promise: A bot in Slack that answers questions from a company's own knowledge.
- Why it died: Slack and knowledge tools such as Guru and Confluence added similar answers themselves, and AI features became cheap.
- Record: https://www.loot-drop.io/startup/1031-niles

### Kite DevTool (2014 to 2024; funding US$21M; cause: Competition)

- Promise: AI code completion inside the programmer's editor.
- Why it died: It never made a free tool pay, and GitHub Copilot arrived with a much stronger model and a large platform behind it.
- Record: https://www.loot-drop.io/startup/2491-kite-devtool

### Neeva (2019 to 2023; funding US$77M; cause: No Market Need)

- Promise: A paid search engine with no ads, sold on privacy.
- Why it died: Too few people paid for search, the cost of each new user was high, and ChatGPT changed what people expected from search.
- Record: https://www.loot-drop.io/startup/2288-neeva

## What changed

Each of the five sold one useful piece of work, writing, saving, answering or searching, and a larger platform gave the same piece away inside a product people already used. Elephas faced the same thing: the Starter Story breakdown mentions a wave of cancellations when ChatGPT became popular, and the 2026 home page compares the app directly with ChatGPT, Claude and NotebookLM.

Two things differ. The model is rented, not built: the user's own key, or open-source models in the company's cloud account, so the team carried no research cost and little usage cost, which the funded predecessors could not say. And the product moved to a job the free chat apps do not do on the computer: keeping confidential files local, hiding personal details before a cloud model sees them, and citing the page of every answer. The co-founder said that this combination, sold mostly through search pages, reached US$150,000 a year with two founders and no investors that any source names.

## Can one person copy Elephas now?

**Verdict: The method, yes. The early timing, no.**

A native app around a public AI API was easy to sell in mid-2022 because almost nothing else existed; today every model maker ships its own desktop app. What can still be copied is the structure: a narrow job for a group that cannot paste its files into a chatbot, no model cost carried by the company, and customers found through community posts first and search pages later.

1. Pick users whose documents must not leave their computer, such as lawyers, accountants or consultants, and build for that rule.
2. Let users bring their own API key or run a local model, so that model usage is never your largest cost.
3. Post short demonstrations of one concrete use in the online communities where those users talk, and answer every reply.
4. Write search pages that answer the questions those users type into Google, and keep them updated.

**What is still open.** Private, local AI for other regulated groups: medical practices, auditors, patent attorneys, and Windows users, whom Elephas does not serve yet.

## What could go wrong for Elephas?

- **What is not verified.** The figure of US$150,000 a year rests on one source: Ayush Chaturvedi's Starter Story interview, published on 21 January 2026, with a recording date that is not stated. The Gumroad and Mac App Store screens shown in it were described by ProvenStartups; this page could not view them. The only other dated revenue figure is US$1,000 a month in November 2022, from Analytics India Magazine. Latka's estimate of US$15 million a year disagrees by a factor of 100. A reader could look for a payment-processor profile (TrustMRR had none for Elephas on 25 October 2026), Setapp or App Store figures, or a second interview with a later date.
- **A new app replaced the one that earned the money.** The reported revenue came from the earlier app, which included writing tools and was sold partly as lifetime licences. In 2026 the company launched a separate app with new accounts and monthly plans, and earlier licences do not carry over. How many customers move is unknown.
- **Free AI products cover much of the same work.** ChatGPT, Claude and Google's NotebookLM read uploaded documents and cite them. Elephas's answer is local storage, hidden personal details and on-computer models. The 2026 app also uses one model chosen by the company for its own AI service, so answer quality depends on open-source models.
- **Marketing aimed at students.** The company blog markets the app to students. A list of AI homework helpers calls Elephas "the ultimate AI-powered homework and study assistant for Mac users". The post on essay apps ends: "Elephas is hands down the best ChatGPT app for essay writing in 2026." The same post also says: "However, submitting AI-generated content as entirely your own work is considered cheating." Aiolum publishes writing tools marketed to students only when the student marketing also says that handing in AI text as one's own is cheating, as this post does. No page read for this story sells AI-detector bypass or "undetectable" text.
- **Search traffic is not owned.** The co-founder said in the interview that search pages brought most of the growth. A change in Google's ranking, or AI answers that keep people on the search page, can reduce that traffic without warning.

## Sources

1. Starter Story on YouTube, 21 January 2026: How I Built a $13K/Month SaaS (interview with Ayush, Elephas) — https://www.youtube.com/watch?v=R4BS_UiTBPw (grade D). Read the watch page: title, date, description and chapter list. The transcript was not read. Source for the US$150,000 a year.
2. ProvenStartups case page: Elephas, $150K/year (founder-reported, captured 26 September 2026) — https://provenstartups.com/projects/elephas-local-ai-knowledge-assistant (grade D). Read the free sections in full. A summary of the Starter Story video. Source for the on-camera Gumroad and App Store figures, the 2023 search revenue, the first US$3,000 a month and the sales channels.
3. Analytics India Magazine, 15 November 2022: Satoshi of AI: Kamban, an India-based AI Writing Tool Developer — https://analyticsindiamag.com/ai-features/meet-satoshi-of-ai-kamban-a-chennai-based-ai-writing-assistant-developer (grade C). Read in full. Founder interview. Source for the US$1,000 in monthly recurring revenue, the Swift code and the user's own OpenAI keys.
4. Ness Labs, 7 February 2024: Augmenting Human Creativity with Ayush Chaturvedi, Co-Founder of Elephas — https://nesslabs.com/elephas-featured-tool (grade C). Read in full. Co-founder interview. Source for the founding account and the Super Brain feature.
5. Website Planet: Elephas Founder Kamban S On Freedom, Friction And Saving Content Costs (undated; republished by Superframeworks on 24 September 2024) — https://www.websiteplanet.com/blog/elephas-interview/ (grade C). Read in full. Founder interview. Source for the team rule, the Reddit reaction and the SEO effort.
6. Hacker News, 9 June 2022: Show HN: Personal AI Writing Assistant for Mac, and later posts by the same account — https://news.ycombinator.com/item?id=31678205 (grade C). Read the post and the founder's replies through the Hacker News search API (hn.algolia.com). Source for the launch date and the user's own OpenAI keys.
7. Elephas home page, read 25 October 2026 — https://elephas.app/ (grade C). Read in full. Source for the product, the three AI routes, the user count and the comparison with ChatGPT, Claude and NotebookLM.
8. Elephas pricing page, read 25 October 2026 — https://elephas.app/pricing (grade C). Read in full. Source for the current plans.
9. Elephas about page, read 25 October 2026 — https://elephas.app/about (grade C). Read in full. Source for the company name and the founding year.
10. Elephas terms of use and privacy policy, last updated October 2026 — https://elephas.app/privacy (grade C). Read both in full (terms at https://elephas.app/terms-of-use). Source for Liberation Softwares LLP, the AI providers, Gumroad and the service providers.
11. Elephas: Using the earlier Elephas Mac app?, read 25 October 2026 — https://elephas.app/classic (grade C). Read in full. Source for the differences between the earlier app and the 2026 app.
12. Elephas pricing page, Wayback Machine copy of 24 June 2025 — https://web.archive.org/web/20250624093807/https://elephas.app/pricing (grade C). Read in full. Source for the 2025 prices, the lifetime licence and the user count.
13. Apple App Store listings of the Elephas apps and the developer page, read 25 October 2026 — https://apps.apple.com/us/app/elephas-secure-ai-for-docs/id6741066105 (grade C). Read the listings of the earlier app and its iOS version, and the developer's other apps (BeeMind, AI Deck). Source for the App Store prices and the providers named.
14. Starter Story: How Elephas Co-Founders Built a Leading AI Writing Assistant for Mac — https://www.starterstory.com/elephas-breakdown (grade C). Read the free part in full. A compiled summary with no revenue figure. Source for the 2023 start of the co-founder and the churn after ChatGPT.
15. Elephas blog posts aimed at students: Top 10 AI Tools for Students (26 December 2024) and Best ChatGPT Apps for Essay Writing (24 October 2024) — https://elephas.app/blog/best-ai-tools-homework (grade C). Read both in full (second at https://elephas.app/blog/best-chatgpt-app-essay-writing). Source for the marketing to students and the content marketer role.
16. Latka company page: Elephas revenue estimate — https://getlatka.com/companies/elephas.app (grade D). Read in full. A third-party estimate that the page itself marks as an estimate.
17. Loot Drop database export, 1,749 failed startups — https://www.loot-drop.io/ (grade C). Local export. AI-written summaries, not independently verified. Source for the failed predecessors.