LiveKit
Open-source real-time audio, video and data infrastructure for AI voice agents — the LiveKit Agents framework, the LiveKit SFU server and LiveKit Cloud ship the low-latency transport layer behind ChatGPT voice mode and a developer base disclosed at 100,000+ at the January 2026 Series C round.
The Business
LiveKit builds the open-source real-time audio, video and data transport layer that AI voice agents run on. The product line spans three main surfaces: the open-source LiveKit server (Apache 2.0, the SFU and signalling stack distributed via the livekit/livekit GitHub repository), the LiveKit Agents framework (the developer-facing Python and Node libraries that wrap OpenAI Realtime, Gemini Live, Claude voice and other realtime model APIs into a deployable agent), and LiveKit Cloud (managed hosted infrastructure with a free developer tier, paid tiered consumption pricing and a self-host enterprise alternative). The company is privately held — founded 2021 in San Francisco by Russ d’Sa and David Zhao, both still in their founding roles in 2026 — and has raised approximately $150M of external capital through the January 2026 $100M Series C at $1.0B post-money, led by Index Ventures with Redpoint Ventures and Hanabi Capital participating.
Customers and Distribution
LiveKit does not file public financials; the primary published commercial signals are the 200,000+ developer base and the “billions of calls/year” usage figure disclosed at the January 2026 Series C announcement, with ChatGPT voice mode named as the marquee customer surface. Distribution sits across four channels: the open-source LiveKit server distributed via the livekit/livekit GitHub repository (Apache 2.0, self-hostable on any infrastructure and the principal driver of the 200,000+ developer count), the LiveKit Agents framework on PyPI and npm (the developer-facing entry point for building voice-AI agents on the OpenAI Realtime API and other realtime model surfaces), LiveKit Cloud managed infrastructure (free developer tier plus tiered consumption pricing for production workloads), and the strategic OpenAI relationship (ChatGPT voice mode running on LiveKit infrastructure, with LiveKit named as a launch partner on the October 2024 OpenAI Realtime API release). Named customer disclosures across the Series C cycle have centred on the OpenAI partnership and the developer-platform scale, with the company declining to publish a stand-alone ARR figure tied to LiveKit Cloud consumption.
Model Strategy
LiveKit is a Verticals-first play under the IM Framework eight-trajectories taxonomy as it applies to real-time voice infrastructure: the strategic bet is that vertical depth on real-time audio and video transport for AI agents — the SFU server, the signalling layer, the Agents framework, the OpenAI Realtime / Gemini Live / Claude voice integrations and the self-host / Cloud optionality — beats horizontal generalist communications-platform plays at managing the voice-agent lifecycle. The LiveKit server is the open-source primitive that anchors developer adoption; the Agents framework converts that adoption into a managed surface where the OpenAI Realtime API and other realtime model APIs are wrapped into deployable agents; LiveKit Cloud converts production deployments into managed-consumption revenue without forcing customers to leave the open-source SDK behind. Above that, LiveKit has consistently invested in the open-source distribution motion — the livekit/livekit repository, the LiveKit Agents 1.0 framework launch in October 2024 alongside the OpenAI Realtime API release, the published developer documentation and the conference cadence — rather than competing as a frontier-model lab or as a managed-only platform. The D4a supplier-diversity sub-rubric was held at 7 in the v1.6 evidence pass on the strength of the multi-cloud deployment posture and the multi-model agent-framework surface (OpenAI Realtime is the marquee integration but not the sole-source dependency). The portability profile is the defining structural variable — open-source self-host is the deliberate strategic feature that drives developer adoption and the same feature caps the share of usage that converts into LiveKit Cloud revenue.
At A Glance
The Numbers
Trend charts are not shown for LiveKit — only single-point data is currently available. See At A Glance above for the most recent disclosed values.
Leadership Team
LiveKit is privately held and does not disclose a full C-suite. Founding duo (d’Sa, Zhao) remain in their roles through the January 2026 Series C, which is a meaningful continuity signal for a five-year-old infrastructure company. CFO and CRO appointments are not separately public; senior engineering and developer-relations hires are visible through the LiveKit blog and conference cadence rather than through formal C-suite announcements. A previously-listed Head of Product entry was removed during fact-checking on 2026-06-02 after verification showed the named individual was not a LiveKit employee.
IM Framework Scoring
IM’s structured assessment of LiveKit’s competitive position. The summary below is the headline; expand “Show the full analyst-grade analysis” near the bottom for the per-dimension reasoning and evidence. Methodology →
Funding History
| Date | Round | Raised | Post-money | Lead investor(s) |
|---|---|---|---|---|
| Jan 2026 | Series C | $100M | $1.0B | Index Ventures |
| Mar 2024 | Series B | $22.5M | — | Index Ventures (with Redpoint Ventures) |
| Aug 2022 | Series A | $22M | — | Redpoint Ventures |
| 2021 | Seed | ~$7M | — | Redpoint Ventures (with Hanabi Capital, individual angels) |
Cumulative external capital approximately $150M through the January 2026 $100M Series C at $1.0B post-money, led by Index Ventures with Redpoint Ventures, Hanabi Capital and other Series A/B participants returning. Round-by-round figures from LiveKit’s own blog and the Series C cycle of named-press coverage (TechCrunch, Bloomberg, SiliconANGLE). The Series C announcement is the canonical reference for the developer-base and OpenAI partnership disclosures cited in this page.
Competitive Landscape
LiveKit’s competitive set sits in three concentric rings: the voice-communications incumbent pivoting into voice-AI agents (Twilio with ConversationRelay), the open-source real-time stacks competing on the same transport-plus-agent-framework surface (Daily with Pipecat, Agora on the WebRTC tier), and the higher-level managed voice-agent platforms competing on the developer-builder surface (Vapi, Retell AI). LiveKit is unusual in the set because the same OpenAI Realtime API that anchors the customer disclosure is simultaneously a dependency (LiveKit Agents wraps the OpenAI surface) and a strategic axis (the relationship is the marquee Series C disclosure).
| Competitor | Positioning | Distribution edge | Threat profile |
|---|---|---|---|
| Twilio (Public (NYSE: TWLO)) |
Long-standing programmable-communications platform (voice, video, messaging) now pivoting hard into voice-AI agents with the ConversationRelay product and the OpenAI Realtime API integration announced through 2025-26. The incumbent in voice channels, working to translate that into agent transport. | Direct enterprise sales plus a developer self-serve channel built over fifteen years; a deep installed base in contact centres, two-factor authentication and customer-communications workflows. | High — the only voice-platform incumbent with comparable developer mindshare; channel breadth and existing customer relationships are the structural threat, partly offset by the heavier legacy stack and slower agent-framework cadence relative to LiveKit Agents. |
| Daily | WebRTC-based real-time video and audio platform with a Pipecat open-source voice-agent framework. The closest direct mirror of LiveKit’s transport + agent-framework combination on the open-source voice-AI side. | Direct developer API; Pipecat open-source distribution; managed Daily Cloud for video and voice workloads. | Medium-high — closest pure-play substitute on the open-source voice-agent stack; smaller installed base than LiveKit at the Series C disclosure but credible on developer-experience and the agent-framework lane. |
| Agora (Public (NASDAQ: API)) |
Real-time engagement platform built on a global WebRTC / SD-RTN network; the long-standing scale incumbent on WebRTC voice and video transport for the consumer-social and live-streaming cohort. | Direct developer API; global network footprint with strong Asia-Pacific presence; SDK distribution at consumer-app scale. | Medium-high — structural competitive pressure on the WebRTC transport layer where Agora’s scale is the comparable, partly offset by Agora’s lighter agent-framework story today relative to LiveKit Agents. |
| Vapi | Voice-agent API platform built on top of underlying transport providers; positioned as the higher-level managed-agent layer with batteries-included LLM / TTS / STT orchestration. Direct competitor to LiveKit Agents on the developer-platform layer for voice-AI builders. | Direct developer API; pay-per-minute pricing; self-serve onboarding aimed at the YC / indie-builder cohort. | Medium — flanking risk on the agent-framework layer for less-sophisticated developers who want a managed pipeline; narrower than LiveKit on the open-source transport layer that anchors the stack. |
| Retell AI | Managed voice-agent platform pitched at contact-centre and customer-experience deployments; bundles transport, LLM orchestration, telephony and analytics into a single managed service. | Direct developer and enterprise channel; vertical-targeted (contact centre, outbound calling); pricing on minutes-consumed. | Medium — narrower vertical lane (contact-centre / phone calling) rather than a horizontal voice-infrastructure substitute; competitive overlap with LiveKit on the higher-level managed-agent surface. |
Pricing benchmark: LiveKit Cloud prices on managed-real-time-minutes consumption (published tiered pricing including a free developer tier) with LiveKit Agents available open-source on self-hosted infrastructure at no marginal Cloud charge. Twilio prices on per-minute voice plus separate ConversationRelay charges; Agora prices on per-minute WebRTC; Vapi and Retell AI price on per-minute managed-agent minutes inclusive of LLM and TTS / STT pass-through. The competitive frame is therefore the open-source transport layer plus the agent-framework developer experience — the LiveKit Agents repo, the OpenAI Realtime integration and the self-host / Cloud optionality — not headline per-minute price alone.
Potential Risks
The case for LiveKit at IM Framework 7.80 rests on the open-source SFU plus LiveKit Agents framework owning the real-time voice transport primitive for AI agents, the 200,000+ developer base disclosed at the January 2026 Series C, the ChatGPT voice mode customer disclosure, and a strategic-investor cohort (Altimeter, Redpoint, Hanabi) that has backed the trajectory across three rounds. The case against splits into five risks of differing magnitude — with OpenAI dependency the most asymmetric, portability across the open-source / Cloud boundary the most structural, and competitive substitution from Twilio and the voice-agent cohort the most active.
OpenAI dependency — marquee customer and strategic asymmetry
OpenAI’s ChatGPT voice mode running on LiveKit is the single load-bearing customer disclosure in the Series C cycle and the canonical voice-AI consumer surface. The flip side is asymmetric: OpenAI is simultaneously a vendor that publishes the Realtime API LiveKit Agents wraps, a major customer of the transport layer, and an entity that could in principle absorb the transport layer in-house. The dependency runs in both directions — LiveKit’s usage scale depends materially on the OpenAI relationship continuing on current terms, and OpenAI’s consumer voice product depends on the LiveKit stack. A renegotiation or in-housing decision by OpenAI would be a single-customer revenue and narrative shock; the offsetting argument is the broader 200,000+ developer base disclosed at the Series C and the multi-model agent-framework posture.
Portability and substitutability — open-source self-host caps Cloud capture
The LiveKit server is open-source (Apache 2.0) and self-hostable, which is a deliberate strategic feature — it anchors developer adoption, drives the 200,000+ developer count and is the moat against managed-only competitors. The flip side is the same feature caps the share of usage that converts into LiveKit Cloud revenue: any sufficiently sophisticated customer can self-host on its own infrastructure and consume the open-source SDK without paying LiveKit. The D1c portability sub-rubric was held at 5 in the v1.6 evidence pass on this basis. The conversion economics from open-source usage to Cloud revenue is the most-watched commercial variable.
Competitive substitution — Twilio, Daily, Agora and the managed-agent cohort
Twilio’s voice-platform pivot (ConversationRelay + OpenAI Realtime integration) is the highest-stakes competitive substitution risk — Twilio brings a fifteen-year installed base in voice channels into the agent transport layer. Daily / Pipecat competes head-on on the open-source voice-agent surface; Agora on the WebRTC transport tier at scale; Vapi and Retell AI on the higher-level managed-agent platform. None has matched LiveKit’s combination of open-source SFU + Agents framework + OpenAI customer disclosure at the Series C cycle, but every named competitor is escalating into the same lane. The substitution dynamic is the principal structural risk on the defensibility composite.
Regulatory exposure — voice-AI consent, recording and impersonation regimes
Real-time voice agents sit inside the voice-recording and two-party-consent regulatory regimes in the US (state-by-state, with two-party-consent states including California, Florida and Pennsylvania) and the broader EU AI Act framing for AI-generated voice content. The D4c regulatory-exposure sub-rubric was held at 5 in the v1.6 evidence pass on this basis. Active variables: voice-cloning and impersonation regulation; outbound-calling and TCPA exposure where LiveKit Agents are used in customer-contact workflows; EU AI Act transparency obligations on AI-generated voice; and the ongoing US state-level legislative cycle on synthetic media. The risk is shared across the voice-agent cohort but is a real and active gating constraint on certain deployment lanes.
Capital position relative to platform peers
The January 2026 $100M Series C at $1.0B post-money (Index Ventures lead, Redpoint Ventures and Hanabi Capital participating) gives LiveKit multi-year runway and was the basis for the D4d sub-rubric upgrade to 8 in the v1.6 evidence pass. Cumulative external capital of approximately $150M is materially smaller than the capital base of Twilio (public, multi-billion revenue), Agora (public) and the better-funded managed-agent peers. The argument is that capital efficiency on the open-source distribution model offsets the absolute-capital gap; the watched event is whether a flat or higher round at the next priced re-mark confirms the trajectory.
Recent IM Coverage
Show recent press coverage of LiveKit
- Jan 2026 — LiveKit raises $100M Series C at $1B valuation — led by Index Ventures, with Redpoint and Hanabi participating.
- Jan 2026 — LiveKit, which powers ChatGPT’s voice mode, raises $100M at a $1B valuation.
- Jan 2026 — LiveKit hits unicorn status with $100M raise as OpenAI partnership powers voice AI surge.
- Oct 2024 — LiveKit Agents 1.0: a framework for building real-time multimodal AI applications.
- Oct 2024 — OpenAI introduces Realtime API — LiveKit named as a launch partner for real-time agent deployments.
- Mar 2024 — LiveKit raises $22.5M Series B led by Index Ventures — doubles down on the real-time WebRTC stack for AI agents.
- Aug 2022 — LiveKit raises $22M Series A led by Redpoint Ventures to build the open-source real-time SDK for the next era of communications.
Curated feed of named-source coverage — LiveKit’s own blog and the named-press cycle around the January 2026 Series C and the OpenAI Realtime API launch (TechCrunch, SiliconANGLE). We exclude PR-wire reposts of the same release, aggregator round-up pieces and subscription-research summaries. The Series C cycle is the primary disclosure anchor for the 200,000+ developer base, the “billions of calls/year” usage figure and the ChatGPT voice mode customer disclosure cited in this page.
Show the source register for the figures on this page
IM operates a primary-source-where-possible discipline. The figures above come from:
- Revenue (basis-disclosure note): LiveKit is private and does not file public financials. The Series C cycle disclosed the 200,000+ developer base and “billions of calls/year” usage figure as the headline commercial signal but did not separately disclose an ARR figure tied to LiveKit Cloud paid consumption. Named-press triangulation has placed LiveKit in the early-stage commercial-traction range typical of an open-source platform at this scale, with the OpenAI ChatGPT voice mode disclosure as the marquee customer reference. We decline-to-publish a precise stand-alone LiveKit revenue number pending a primary disclosure from the company or a named-press article anchored to a primary source. Reference: LiveKit Series C announcement.
- Usage — developer base and call volume: LiveKit’s January 2026 Series C announcement disclosed a 200,000+ developer base on the platform and “billions of calls/year” running through the LiveKit infrastructure, with ChatGPT voice mode named as the marquee customer surface. The livekit/livekit GitHub repository is the canonical open-source distribution point for the SFU server and SDKs; the agents framework is at livekit/agents. We use developer count and disclosed call volume as the headline usage benchmark rather than monthly active developer counts, which are not company-disclosed.
- Headcount (basis-disclosure note): LiveKit is private and does not separately disclose headcount in a formal filing. Public-interview commentary from Russ d’Sa and named-press coverage have placed the company in the small-team range typical of a Series B-to-C open-source infrastructure platform — roughly 50-150 employees through the January 2026 Series C cycle, with continued hiring implied but not company-disclosed at a precise headcount. We decline-to-publish a precise headcount and reference the LiveKit careers page as the canonical entry point, with named-press triangulation rather than third-party trackers as the source-of-record for any cited figure.
- Funding to date: Cumulative external capital approximately $150M through the January 2026 $100M Series C at $1.0B post-money, led by Index Ventures with Redpoint Ventures and Hanabi Capital participating — per LiveKit’s own announcement and corroborated by TechCrunch and SiliconANGLE. Prior rounds (Series A $22M August 2022 led by Redpoint, Series B $22.5M March 2024 led by Index Ventures) from LiveKit’s own blog cycle.
Methodology & Disclaimer
For metric definitions, source-tier hierarchy, and decline-to-publish rules, see the tracker methodology. Confidence dots (• green / • amber / • red) follow the same convention as the AI Tracker.
Spotted a figure you believe is wrong? Send corrections to info@informationmatters.net.
Information Matters Framework scores are the considered opinion of the IM team — human and AI — applied to publicly-available evidence under a disclosed methodology. They are not statements of fact about the companies scored and they are not investment advice.
