Speak · 17× · naposledy 30. 6. 2026
Speak (by Speakeasy Labs, San Francisco) is an AI-powered language learning app founded in 2016 that currently teaches English, Spanish, French, Korean, Japanese, and Italian. Its core feature is the "Speak Tutor," an AI conversation partner built on OpenAI's GPT-4 along with the Realtime API and Whisper for speech recognition, providing feedback on pronunciation, grammar, and vocabulary. The app is available as a mobile app (iOS/Android) and web app, offering two subscription tiers (Premium and Premium Plus) alongside a 7-day free trial. Speak is backed by the OpenAI Startup Fund and reached a $1 billion valuation in late 2024.
Vlastnosti
| Base Model | OpenAI GPT-4 (AI Tutor) plus OpenAI Realtime API for live conversation/roleplay |
| Integrations | OpenAI GPT-4, OpenAI Realtime API, Whisper API (speech recognition); former ChatGPT plugin integration |
| Collaboration | Single-user app; plus enterprise offering "Speak for Business" for organizations |
| Platform | iOS app, Android app (Google Play), and web app (app.speak.com) |
| Price | Premium approx. $20/month or ~$99/year; Premium Plus higher-priced (price varies by market) |
| Release Date | App available since 2016 (Speakeasy Labs, San Francisco); originally an English-learning app for South Korea |
Další produkty v této kategorii: AI výuka a produktivita
Zdroje (17)
Company Analysis: Speak
As of 22/08/2026On a risk‑adjusted basis for a private, late‑stage growth investment, the opportunity set modestly outweighs the risks. Speak combines (1) a clearly differentiated AI‑native product focused on spoken fluency, (2) strong engagement growth and evidence of habit formation, (3) a solid capital base and backing from top‑tier investors, and (4) a credible roadmap for geographic expansion beyond its Asian strongholds. While competition is intense and the unicorn valuation limits margin of safety, the company’s specialization and early mover advantage in AI speaking‑coaching give it a realistic path to build a premium, high‑ARPU niche. For investors with access to private rounds and tolerance for illiquidity, this supports a cautiously positive BUY stance in EUR terms, contingent on continued execution and disciplined pricing/monetization.
Key Takeaways
- Speak is a privately held, AI‑native language‑learning app (no listed equity, no EUR quote) focused on spoken practice with an AI tutor; it has become the #1 paid education app in core Asian markets (Korea, Japan, Taiwan) and surpassed 10M downloads globally, positioning it as a leading challenger to Duolingo in speaking‑centric learning. (speak.com)
- The company has raised roughly $150–160m in total funding and reached unicorn status (~$1B valuation) following a $78m Series C led by Accel and the OpenAI Startup Fund; it remains well‑capitalized for global expansion but with typical late‑stage growth expectations from investors. (speak.com)
- Engagement and usage metrics are strong: in 2025 users logged 19.6m learning hours in‑app, up ~85% YoY, indicating rapid adoption and habit formation ahead of planned expansion into Europe, Latin America and the US in 2025–26. (speak.com)
- The competitive landscape is intense, with Duolingo, Babbel and newer AI‑first apps (e.g., Enverson, Praktika, Loora) all pushing speaking‑oriented features; user reviews praise Speak’s design and focus on conversation but highlight high pricing and aggressive paywalls as friction points. (forbes.com)
- Traffic and awareness indicators show solid but not explosive momentum: speak.com’s organic Google traffic was ~119k visitors in July 2026, and app‑store ratings are high, but Google Trends‑style interest appears to have peaked around the late‑2024 unicorn/Series‑C news and has since normalized, suggesting the brand must now sustain growth through execution rather than hype. (sitestatsdb.com)
Action Ideas
For investors able to access private growth‑stage deals (VC/PE or secondary), Speak offers an attractive AI‑native consumer edtech play: it combines strong product–market fit in speaking‑first language learning, high engagement growth (19.6m hours in 2025, +85% YoY), and a credible path to global expansion beyond its Asian beachhead. Backing from Accel, OpenAI Startup Fund and other tier‑one investors, plus unicorn‑level valuation, indicates institutional confidence in its technology moat and monetization potential. If management executes on Europe/LatAm/US rollout and maintains retention, revenue could scale meaningfully over the next 3–5 years, supporting upside from current private valuations in EUR terms.
Horizon: 48 mo.
Existing late‑stage private holders who can realize gains may consider trimming or exiting: Speak already commands a unicorn valuation with no clear public‑market liquidity path, while competitive intensity and AI commoditization are rising. Duolingo is scaling AI features across a massive user base, and multiple AI‑first rivals are emerging with similar speaking‑coach propositions. Google Trends‑style interest and web traffic suggest normalization after the 2024 funding spike rather than accelerating brand momentum. If revenue growth decelerates as acquisition costs rise and pricing remains a friction point, current valuations could prove full, leaving limited multiple expansion from here.
Horizon: 18 mo.
For investors already committed and unable to easily rebalance, maintaining exposure while closely monitoring operating KPIs is reasonable. Speak’s core metrics (engagement hours, subscriber base, geographic expansion) and unit economics will be decisive over the next 12–24 months. A HOLD stance allows participation in potential upside from successful Western expansion and improved AI‑driven personalization, while avoiding fresh capital deployment at a valuation that already prices in substantial growth. Investors should seek quarterly or annual updates on revenue growth, churn, CAC/LTV, and progress in Europe/LatAm/US before revising to a more decisive BUY or SELL.
Horizon: 24 mo.
Google Trends · → stable
Proxy indicators for Google Trends—such as speak.com’s organic search traffic (~118k Google visitors in July 2026) and the timing of major newsflow—suggest that global search interest in Speak rose sharply around its late‑2024 unicorn/Series‑C announcement and 2025 media coverage, then normalized into a relatively stable plateau through mid‑2026. Interest no longer appears to be accelerating, but it also has not collapsed; instead, it reflects a maturing brand moving from hype‑driven spikes to steady, product‑led awareness as the app expands beyond its initial Asian markets.
Contrarian Insights
- • Despite the crowded AI language‑learning space, Speak’s deep specialization in spoken interaction and its early, dense foothold in Korea and other Asian markets may give it a defensible niche that generalist competitors underestimate. Rather than needing to beat Duolingo globally, Speak could build a highly profitable, premium segment focused on serious speaking outcomes, with ARPU and retention strong enough to justify its valuation even at a smaller absolute user base.
- • The current narrative focuses on consumer subscriptions, but Speak’s technology stack (speech recognition tuned to foreign‑accented speakers, scenario‑based AI coaching) could be more valuable in B2B/B2G contexts—corporate training, test prep, or public‑sector language programs—than in the crowded direct‑to‑consumer app market. If management pivots to or layers in enterprise contracts, revenue quality and margins could improve, a scenario not fully reflected in today’s consumer‑centric expectations.