Captions (app)
Last reviewed
Jun 4, 2026
Sources
21 citations
Review status
Source-backed
Revision
v1 ยท 2,178 words
Improve this article
Add missing citations, update stale details, or suggest a clearer explanation.
Suggest edit
CtrlK
Last reviewed
Jun 4, 2026
Sources
21 citations
Review status
Source-backed
Revision
v1 ยท 2,178 words
Add missing citations, update stale details, or suggest a clearer explanation.
Captions (stylized captions.ai, legally NOCAP, Inc.) is an artificial intelligence-powered video creation and editing company and product based in New York City. Founded in 2021 by Gaurav Misra and Dwight Churchill, the company first became known for a mobile app that automatically generated subtitles for "talking videos" and then corrected a speaker's gaze toward the camera with a feature called AI Eye Contact. It later expanded into a broader AI creative studio offering automated editing, dubbing and translation, and AI-generated avatars, and in 2025 introduced an in-house video-generation foundation model named Mirage. In September 2025 the parent company rebranded itself from Captions to Mirage to reflect a shift toward building proprietary video models, while keeping the Captions name on its consumer app. By 2024 the product had surpassed 10 million mobile downloads, and the company had raised roughly $100 million in venture funding at a reported $500 million valuation, followed by a $75 million growth round in 2026.
Captions was founded in 2021 in New York City. Co-founder and chief executive Gaurav Misra previously led design engineering at Snap, where he worked on social products including Spotlight and Snap Map, and earlier was a software engineer at Microsoft. Co-founder Dwight Churchill, who serves as chief operating officer, was previously an early member of the team that built Marcus, Goldman Sachs' consumer finance business. The legal entity operates as NOCAP, Inc., doing business as Captions.
According to Misra, the product found its market almost by accident. The team had built a video app that was not gaining traction, put a roughly $10-per-month paywall on it, and largely moved on. Returning months later, they discovered that paying users had accumulated and were specifically using the captioning feature, which led the founders to refocus the entire company on AI-assisted video for creators. The app concentrated on the "talking video" format, in which a creator addresses the camera directly to share opinions, advice, or personal stories, a format Misra had watched rise during his time at Snap.
Captions raised across several rounds, with early backing from Sequoia Capital and Andreessen Horowitz (a16z) and later rounds led by Kleiner Perkins and Index Ventures. Investors and aggregators describe an early seed investment followed by a Series A co-led by Sequoia and a16z in 2022. The company's Series B, announced on June 22, 2023, raised $25 million led by Kleiner Perkins with participation from Sequoia Capital, Andreessen Horowitz, and SV Angel, bringing total capital raised to about $40 million at that point.
On July 9, 2024, Captions announced a $60 million Series C led by Index Ventures, which valued the company at $500 million and brought total funding to roughly $100 million. Returning investors Kleiner Perkins, Andreessen Horowitz, and Sequoia Capital took part, alongside new investors Adobe Ventures, HubSpot Ventures, and the actor Jared Leto. The company framed the round as a commitment to "invest $100M into advancing generative video research from New York City," with the money going toward expanding its machine-learning team and technical infrastructure.
On March 24, 2026, by then operating as Mirage, the company announced $75 million in growth financing from General Catalyst's Customer Value Fund, capital structured to be repaid out of revenue rather than diluting equity in a traditional priced round.
| Round | Date | Amount | Lead investor | Notable participants | Reported valuation |
|---|---|---|---|---|---|
| Seed / early | 2021-2022 | early-stage | Sequoia Capital, Andreessen Horowitz | SV Angel | not disclosed |
| Series A | 2022 | undisclosed | Sequoia Capital and Andreessen Horowitz (co-led) | various | not disclosed |
| Series B | June 22, 2023 | $25 million | Kleiner Perkins | Sequoia, a16z, SV Angel | not disclosed (total raised ~$40M) |
| Series C | July 9, 2024 | $60 million | Index Ventures | Kleiner Perkins, a16z, Sequoia, Adobe Ventures, HubSpot Ventures, Jared Leto | $500 million (total ~$100M) |
| Growth (as Mirage) | March 24, 2026 | $75 million | General Catalyst (Customer Value Fund) | revenue-based financing | not disclosed |
Because some early-stage figures appear only in third-party funding trackers and are reported inconsistently (one source describes a roughly $15 million seed, others list a smaller seed plus a Series A of around $11 million), the precise amounts and dates of the seed and Series A are treated here as approximate; the $25 million Series B (June 2023), the $60 million Series C (July 2024), and the $75 million growth round (March 2026) are the figures confirmed across multiple independent reports.
On November 13, 2024, Captions made its first acquisition, buying AlpacaML (also styled Alpaca), a generative-AI rendering tool founded in 2022 that turned sketches, thumbnails, and images into finished, styled visuals. AlpacaML's chief executive William Buchwalter joined Captions as a research engineer, and job offers were extended to all six of AlpacaML's employees. Around the same time, the company hired Drew Jaegle, formerly of Google DeepMind, as Head of AI; at DeepMind he had worked on multimodal representation and generative models. The acquisition and hire were positioned as the build-out of an in-house AI research arm.
The Captions mobile app, launched for iOS in 2021 and later on Android, is an AI-assisted studio for short-form, creator-style video. Its features grew over time and have included:
The app monetizes through subscriptions; in January 2025 the company moved its mobile app to a freemium model. Earlier reporting cited a subscription tier in the range of roughly $10 per month.
In October 2023, Captions released Lipdub, a free standalone iOS app for AI dubbing and translation. Lipdub translated a video of a single person speaking (up to about one minute) into 28 languages, including French, Hindi, Spanish, Italian, Portuguese, and Japanese, and adjusted the speaker's lip movements to match the target language. It also offered novelty "languages" such as Gen Z slang and pirate speech. Contemporary coverage noted occasional lag between audio and lip movement. The same lip-modification technology underpinned an in-app "AI Lipdub" feature that could change a speaker's mouth movements when the transcript was edited in post-production.
On March 12, 2025, the company introduced Mirage, which it described as the world's first video foundation model purpose-built for generating UGC-style ads and talking content, developed entirely in-house. Rather than applying dubbing or lip-sync to licensed footage of real actors, Mirage generates complete scenes from scratch: hyper-realistic talking people, objects, and backgrounds that do not exist. The company's argument for the approach is that much of human communication is carried by facial expressions, micro-reactions, and body language, so it generates full facial and upper-body motion rather than only matching lips to audio.
Mirage can produce video from a text prompt, a script, a video file, or an audio file alone, and lets users specify a speaker's apparent age, gender, clothing, background, and on-screen products. The company says it supports more than 29 languages and preserves accents in generated speech. Mirage is a generative media model in the same broad family as other diffusion-model and text-to-speech systems, combining synthetic video with synthetic voice; the company has emphasized building the underlying machine learning and deep learning stack itself rather than licensing third-party models.
Mirage Studio, launched in June 2025, is a web platform aimed at brands and advertisers that uses the Mirage model to generate short ads without human talent, stock footage, voice cloning, or traditional lip-syncing. A user can submit an audio file and the system generates the video, including an AI background and AI avatars, or upload selfies to create an avatar of a specific likeness. The company says these AI actors can laugh, sing, rap, gesture, flinch, and convey micro-expressions. Mirage Studio is offered under a business plan reported at $399 per month for 8,000 credits, with a 50 percent discount on the first month for new users. Following the 2026 growth round, the company also described a web-based "marketing suite" for bulk creation and distribution of videos by companies.
On September 4, 2025, the company announced that it was rebranding from Captions to Mirage at the corporate level, while the consumer app would keep the Captions name. Misra framed the change as a move from being a creator-tools company toward being an AI research lab focused on multimodal foundation models for short-form video on platforms such as TikTok, Reels, and Shorts, telling TechCrunch that "the real race for AI video hasn't begun." Under the new structure, the Captions app, the Mirage model, and Mirage Studio all sit under the Mirage brand.
Because the rebrand consolidated AI-generated talking-head video, the company addressed concerns about deepfakes and impersonation. It said it had moderation measures intended to prevent impersonation and to require consent for the use of a person's likeness, while acknowledging that product design "isn't a catch-all" and arguing that part of the answer is a broader "new kind of media literacy."
Captions grew quickly as a consumer product. By the time of its June 2023 Series B, the company reported more than three million users; an October 2023 account cited over 100,000 daily active users producing on the order of a million videos a month. At the July 2024 Series C, the company reported more than 10 million mobile downloads and roughly three million monthly active users.
By the 2026 growth round, the company (as Mirage) reported that more than 200 million videos had been created with its tools to date. Third-party app analytics cited at that time indicated over 3.2 million downloads and about $28.4 million in in-app revenue over the preceding 365 days. The company also said its user base was heavily international, with only about a quarter of revenue coming from the United States.