Speechly provides an API that performs speech recognition and language understanding directly on user devices, reducing costs for applications that need to analyze spoken interactions in real time.
Also known as Speechgrinder
Founders & leadership· Y Combinator alumni (W22)
Speechly was founded in 2016 by Otto Soderlund and Hannes Heikinheimo.
Investors · 5
Company profile
researched Aug 2026Speechly was a Finnish speech technology company, founded in 2016 and headquartered in Helsinki, that provided real-time automatic speech recognition (ASR) and natural language understanding (NLU) through a single API. Its low-latency, fully streaming Spoken Language Understanding offering converted speech directly into actionable intents, which the company positioned as a way to build multimodal voice interfaces where voice input works alongside touch and visual interaction. Deployment options included cloud, on-premise and on-device, and the company offered custom models for specific domains and scalability for hundreds of thousands of hours of audio.
The product was distributed as SDKs and client libraries for the major platforms — JavaScript and React on the web, plus iOS, Android and Unity — along with pre-built UI components, a developer dashboard, a CLI and gRPC protocol definitions. Public demos illustrated two main use cases: real-time transcription of live and pre-recorded audio, and content moderation that highlighted profanity and labelled utterances as offensive or non-offensive in real time. Voice chat moderation for gaming and metaverse audio became the company's most prominent application and the stated rationale for its acquisition by Roblox in September 2023.
Founding story
Speechly was founded in 2016 by a team of scientists, developers and entrepreneurs on the premise that voice would become a primary mode of online interaction. The founding group comprised CEO Otto Söderlund, CTO Hannes Heikinheimo, Chief Scientist Janne Pylkkönen and Chief Architect Markus Lång, several of whom had previously worked on voice products including Apple Siri and Amazon Alexa. Söderlund had earlier founded and sold a Nordic digital management consultancy, while Heikinheimo holds a PhD in machine learning from Aalto University and had worked on Apple Siri, Angry Birds, Nokia Music and Google Maps.
Business model
Speechly sold developer-facing speech technology delivered as an API with accompanying SDKs and UI components, offering flexible deployment options across cloud, on-premise and on-device, together with custom models built for a customer's domain. Developers created and managed applications through a Speechly dashboard and CLI, and integrated the service using client libraries for JavaScript, React, iOS, Android and Unity.
Revenue came from developer and enterprise use of its speech recognition and language understanding APIs and deployable models.
Traction
The company reported a team of about 10-12 people. Its open-source web client monorepo accumulated 185 GitHub stars and 18 forks across 860 commits, and published browser and React client packages on npm alongside Android, iOS and Unity clients.
Latest developments
Roblox acquired Speechly in September 2023 to apply its speech recognition and moderation technology to real-time moderation of voice chat on a platform reporting 65.5 million daily active users; the terms were not disclosed. Following the acquisition, Speechly's public web client repository was archived as read-only in January 2025.
▸Full profile — market position, technology, go-to-market, geography, history
Market position
Speechly competed with the voice platforms of Google, Amazon and Apple by targeting developers who wanted voice capability embedded inside their own products rather than inside a third-party assistant ecosystem; an investor characterized the ambition as being "the Twilio for voice UIs." At the time of its 2019 seed round the company cited a CB Insights forecast of a $49 billion voice recognition market by 2024. It was acquired by a publicly listed platform company in 2023.
The company positioned its combination of natural language understanding with speech recognition in a single streaming pipeline as the source of a faster and more intuitive experience than conventional voice assistant platforms, which it argued handle complex requests inconsistently and lock developers into their ecosystems. Speechly emphasized that customers retained their user data and were not required to monetize it, that integration was possible with existing apps and interfaces, and that on-device processing made large-scale analysis of conversational audio cost-efficient. Its founding team's prior work on Apple Siri, Amazon Alexa and other large-scale products was cited as a technical credential.
Technology
The core technology combined automatic speech recognition and natural language understanding into a single low-latency streaming pipeline that produced transcripts and structured intents as the user speaks, described by the company as Spoken Language Understanding or speech-to-intent. Models could run on-device, on-premise or in the cloud, and could be customized per domain; on-device processing was presented as the mechanism for making large-volume conversational audio analysis affordable. Real-time moderation models detected profanity, abuse and toxicity in audio streams so unacceptable content could be filtered instantly.
Go-to-market
Distribution was developer-led: open-source client libraries published on npm and GitHub, hosted demos, quick-start guides and reference documentation, a self-service dashboard for creating and managing applications, and community support via GitHub Discussions and Issues.
Developers and product teams adding voice interfaces to applications, websites and hardware, with use cases spanning e-commerce voice search and navigation, gaming voice chat and commands, digital health, search, enterprise hands-free workflows, and voice chat moderation for gaming and metaverse platforms.
Geography
Headquartered in Helsinki, Finland, with a developer-oriented product distributed globally through APIs, SDKs and package registries.
History
The company was founded in 2016 in Helsinki, initially operating under the name Speechgrinder before rebranding to Speechly in 2019. In December 2019 it raised a seed round of EUR 2 million (about US$2.2 million) led by Cherry Ventures, with Seedcamp and others participating, to launch its public APIs. It took part in Y Combinator's Winter 2022 batch. In September 2023 Roblox acquired the company for an undisclosed sum, with co-founder and CTO Hannes Heikinheimo joining Roblox as a senior engineering manager. The company's public web client monorepo on GitHub was archived and set to read-only in January 2025.
Compiled by commissioned research from 8 cited public sources — announcements, filings, and press listed under research sources below.
Key figures
latest reportedCompany-reported or press-reported figures, each dated to when it was claimed — not independently audited.
Timeline · 5
launches, deals, and filingsThe speechly/speechly monorepo containing browser and React client libraries, demos and examples was archived and made read-only.
Roblox (NYSE: RBLX) acquired Speechly for an undisclosed sum to strengthen real-time voice chat moderation on its platform. Speechly co-founder and CTO Hannes Heikinheimo became a senior engineering manager at Roblox.
Seed funding led by Cherry Ventures with participation from Seedcamp and others, earmarked for developing and launching Speechly's public APIs.
$2.2M source ↗
The company, previously known as Speechgrinder, rebranded as Speechly in 2019.
Dated company events from announcements, filings, and press; legal rows summarize public dockets and regulator releases.
In the news
▸Research sources · 8
primary sources listed
- Speechly - Stop typing, just speakspeechly.io · web
8 public sources were cited for this profile; the first-party ones are listed here.
Frequently asked questions
- What does Speechly do?
- Finnish developer of real-time speech recognition and natural language understanding APIs, acquired by Roblox in 2023.
- Who founded Speechly?
- Speechly was founded by Otto Soderlund, Hannes Heikinheimo in 2016.
- Who are Speechly's investors?
- Speechly's investors include Cherry Ventures, Seedcamp, SNÖ Ventures, Tq Ventures Llc, Y Combinator.
- Who acquired Speechly?
- Speechly was acquired by Roblox.
- Where is Speechly headquartered?
- Speechly is headquartered in Helsinki, FI.