Carahsoft

Technology

Deepgram brings its Voice AI Platform to the Dell AI Factory with NVIDIA, giving enterprises a path to run speech-to-text, text-to-speech, and voice agents on Dell infrastructure inside their own data centers. The stack pairs Deepgram's models with NVIDIA accelerated computing on Dell servers, operated through the Dell AI Factory model. For security-sensitive workloads, Fortanix Confidential AI adds model-in-use protection on NVIDIA Confidential Computing-enabled GPUs, keeping audio data and model weights encrypted throughout processing. This gives regulated and data-sovereign organizations a supported route to production voice AI without sending audio to a public cloud. A joint reference architecture is in development so teams can deploy a validated configuration rather than assembling one from scratch. If you are planning an on-premises voice AI deployment on Dell infrastructure, reach out via deepgram.com/contact-us to discuss deployment options. Outlinks & Resources Dell AI Factory Deepgram × Fortanix: Private Voice AI for Regulated Industries Deepgram Delivers Private Voice AI to Regulated Industries (June 2026 press release) Contact Deepgram

Technology

Communications / CPaaS

Cloud Infrastructure

Speech to Text

Global

Deepgram and Fortanix bring private voice AI to regulated industries. The partnership, announced in June 2026, lets enterprises run Deepgram's voice AI models entirely within their own environment while their most sensitive data stays protected. The deployment runs Deepgram's models with Fortanix Confidential AI on NVIDIA Confidential Computing-enabled GPUs. Workloads execute inside a hardware-isolated trusted execution environment, a protection traditional infrastructure cannot provide. Audio data and Deepgram's proprietary model weights remain encrypted throughout active use, which protects customer data on one side and guards the models against theft or inappropriate use on the other. For healthcare systems, financial institutions, and public-sector organizations, this closes the gap between voice AI adoption and security mandates. Deployments can meet HIPAA, GDPR, and national data-residency requirements without compromising on model quality. In Deepgram's benchmark, Nova-3 delivers the lowest word error rate of any model tested, and real-time transcription runs at sub-300ms latency inside the customer's own environment. If you are evaluating voice AI for a regulated or data-sovereign environment, reach out via deepgram.com/contact-us to discuss on-premises deployment options. Outlinks & Resources Deepgram Delivers Private Voice AI to Regulated Industries with On-Premises Deployments Powered by Fortanix Confidential AI and NVIDIA Confidential Computing Fortanix Confidential AI NVIDIA Confidential Computing Contact Deepgram

Technology

Vapi is one of the leading voice agent platforms, and Deepgram provides the speech layer inside it. Builders who pick Vapi to ship a production voice agent get Deepgram's Nova-3 STT, Flux for conversational turn detection, and Aura-2 TTS as native model options, configured in a few clicks rather than a custom integration. Vapi handles 62M+ calls monthly with 99.99% uptime, making it the proven choice for enterprises requiring scale across voice, SMS, and chat. Deepgram is integrated for its speed and human-like voice quality, and powers the STT and TTS surface that real production voice agents depend on. Voice agent quality lives in the speech layer. The transcription needs to handle interruptions, background noise, accents, and the alphanumeric read-back that real conversations include. The TTS needs to sound natural at low enough latency that the caller does not notice. Deepgram's voice-native foundational models are built specifically for those requirements, and Vapi exposes them as the platform's recommended models for production agents. The pattern: a developer picks Vapi to handle orchestration (transport, interruption, tool calls, state), picks Deepgram models for STT and TTS, drops in an LLM, and ships a working voice agent in hours. For latency-critical or regulated workloads, the same Vapi configuration runs against Deepgram's self-hosted runtime, keeping audio inside the customer's environment. If you are building a voice agent on Vapi and want production-grade speech, Deepgram is the default. Resources below. Outlinks & Resources Vapi partner page Vapi: Voice AI Platform for Developers (Voice AI Apps) Top Voice AI Agents for 2026: The Ultimate Buyer's Guide Vapi platform Deepgram Voice Agent API Contact Deepgram

Development

Contact Center

Healthcare

Speech to Text

Audio Intelligence

Global

JetBridge is a custom software engineering firm founded by the engineers behind Five9 and DoctorBase. The team builds production-grade voice AI applications for VC-backed startups and mid-market companies, with deep experience in contact center and healthcare environments where accuracy and compliance requirements leave no margin for a mediocre speech layer. JetBridge integrates Deepgram's models as core components in the applications it builds for clients. Nova-3 handles real-time transcription for contact center and CX workflows; Nova-3 Medical covers clinical documentation and healthcare applications where domain-specific vocabulary and HIPAA compliance are requirements, not optional. The team has applied this stack for companies backed by Andreessen Horowitz, including Asimov, and for enterprise clients including LabCorp and Compass. JetBridge runs on a Forward Deployed Engineering model: clients see a working proof of concept before committing to a full engagement. That approach compresses the timeline between architecture review and working software, which matters for startups building toward a fundraise and for mid-market companies running a competitive evaluation. There is no upfront cost to validate the approach before the partnership begins. Teams evaluating Deepgram for a custom integration can reach JetBridge directly at deepgram@jetbridge.com for a free architecture consultation, including cost and timeline estimates. JetBridge responds within minutes. Outlinks & Resources JetBridge: https://jetbridge.com Contact JetBridge: deepgram@jetbridge.com Deepgram Enterprise: https://deepgram.com/enterprise Deepgram Documentation: https://developers.deepgram.com

Channel Partner

Contact Center

Healthcare

Financial Services

Professional Services

Text to Speech

Voice Agent

TTS Flux

Global

Cognizant is one of the world's largest IT services and consulting firms, with more than 300,000 professionals delivering AI, cloud, and digital transformation programs across every major industry. The company integrates Deepgram's voice AI models into the enterprise AI solutions it architects and deploys for its clients. Cognizant's AI practice draws on Deepgram for the speech layer in AI-driven applications: real-time contact center transcription, voice-enabled AI agents, compliance call recording, and speech interfaces for enterprise workflows. Nova-3 handles high-accuracy streaming transcription across Cognizant's core verticals. Flux handles conversational STT in voice agent deployments where built-in turn detection and interruption handling are requirements, not add-ons. The industries Cognizant serves are where voice AI carries the most operational weight. Financial services clients need accurate, auditable call transcription feeding QA and risk systems. Healthcare clients need clinical documentation that meets HIPAA requirements and handles domain-specific vocabulary. Contact center clients need AI agents that can sustain real conversations at volume. Cognizant's Neuro AI platform and AI-led modernization practice provide the implementation scale to move voice AI from proof of concept to production across these environments. Enterprises looking to add voice AI capabilities to a broader AI transformation program can engage Cognizant through their standard advisory process or contact Deepgram's enterprise team to explore the partnership. Outlinks & Resources Cognizant AI Services: https://www.cognizant.com/us/en/services/ai Cognizant Neuro AI Platform: https://www.cognizant.com/us/en/services/ai/enterprise-ai Deepgram Enterprise: https://deepgram.com/enterprise Deepgram Documentation: https://developers.deepgram.com Contact Deepgram: https://deepgram.com/contact-us

Technology

Communications / CPaaS

Speech to Text

Text to Speech

STT Nova

TTS Flux

Deepgram and LiveKit enable developers to build production-ready voice AI agents with real-time speech recognition and natural text-to-speech, all within LiveKit's open-source framework for voice, video, and physical AI agents. LiveKit's Agents framework ships native Deepgram plugins for both STT and TTS. Developers run Nova-3 for high-accuracy streaming transcription across 30+ languages, or Flux for conversational speech recognition with built-in end-of-turn detection and interruption handling. Flux is purpose-built for voice agent workflows where handling barge-in cleanly is the difference between a demo and a production deployment. Model selection is a single configuration line in Python or TypeScript. The integration covers the core voice agent use cases: customer support automation, multilingual voice assistants, meeting bots, and inbound/outbound telephony via LiveKit's SIP integration. Deepgram models are also available through LiveKit Inference, with billing and model access managed directly through LiveKit's platform. LiveKit runs on a global cloud network built for production workloads, with over 300,000 developers and billions of calls processed annually. With voice-native infrastructure and Deepgram's sub-300ms latency speech models, development teams get the fastest path from prototype to production for agents that hold real conversations at scale. Outlinks & Resources LiveKit Agents Framework (Docs): https://docs.livekit.io/agents/ LiveKit Agents Framework (GitHub): https://github.com/livekit/agents LiveKit Deepgram STT Plugin Guide: https://docs.livekit.io/agents/models/stt/deepgram/ LiveKit Deepgram TTS Plugin Guide: https://docs.livekit.io/agents/integrations/tts/deepgram/ LiveKit Deepgram Integration Guide: https://developers.deepgram.com/docs/livekit-integration Build a Voice Agent with LiveKit and Deepgram: https://developers.deepgram.com/docs/build-voice-agent-with-livekit-and-deepgram

Accenture is one of the largest global systems integrators, with deep practices across contact center transformation, customer experience, enterprise generative AI, and public sector technology delivery. Deepgram provides the speech-to-text and text-to-speech layer underneath the voice AI workloads Accenture's clients are building. For enterprises evaluating a partner to bring voice AI into production, the practical questions are about workflow design, integration with existing telephony and CCaaS stacks, and day-two operations. Accenture brings the strategy, integration, and managed-services capacity for that work. Deepgram brings the speech models: Nova-3 for streaming STT, Aura-2 for TTS, and a self-hosted runtime for regulated workloads where audio cannot leave the customer's environment. The combined offering covers contact center modernization (live transcription, agent assist, automated quality management, post-call analytics), enterprise voice agents (real-time STT, conversational turn detection, low-latency TTS), and regulated-industry deployments (healthcare, financial services, public sector) where compliance posture and data residency are required from day one. If you are working with Accenture and evaluating Deepgram, or building a voice AI program and want an implementation partner who can ship into production, deepgram.com/contact-us is the direct path. Outlinks & Resources How Enterprises Can Unlock Voice AI Through the Deepgram Partner Ecosystem Deepgram Self-Hosted overview Deepgram documentation Contact Deepgram

500 Global is one of the largest seed-stage VC firms and accelerators, with roughly $2.3B AUM, more than 2,500 portfolio companies across 80+ countries, and 38+ unicorns. Deepgram is now on 500 FounderHub, the perks platform 500 Global runs for its portfolio. For 500 portfolio companies building voice products (AI agents, contact center software, real-time transcription, voice features in consumer or vertical SaaS), Deepgram provides the speech layer underneath. Nova-3 for streaming speech-to-text, Aura-2 for text-to-speech, and the Voice Agent API for full agent workflows are available through the FounderHub benefit. If you are a 500 Global founder, the path to claim Deepgram credits runs through your FounderHub login. For terms beyond the program tier or for enterprise procurement, deepgram.com/contact-us is the direct route. Outlinks & Resources 500 Global 500 FounderHub Deepgram Startup Program Deepgram documentation Contact Deepgram

Technology

8x8's XCaaS (Experience Communications as a Service) brings cloud contact center, voice, video, chat, and APIs together in a single integrated offering, and Deepgram provides the speech-to-text layer inside it. Real-time transcription, call analytics, and voice intelligence inside 8x8 run on Deepgram's Nova-3 model. Speech accuracy has to hold up in the conditions an 8x8 customer actually runs: diverse accents, background noise, and the domain-specific vocabulary (account numbers, policy IDs, product SKUs) that drives self-service interactions. Nova-3 is built for that envelope, with low enough latency to support agent assist on live calls and clean enough transcripts to feed downstream quality-management and analytics workflows. For enterprises running 8x8 Contact Center, the integration covers the surfaces that depend on speech: live transcription for agent assist, automated QM scoring, and post-call conversational analytics. 8x8 handles the call routing, workforce engagement, and CX surface; Deepgram provides speech recognition designed for real-time production use. If you are running 8x8 and evaluating speech recognition, deepgram.com/contact-us is the direct path. Outlinks & Resources 8x8 Platform 8x8 Contact Center Deepgram documentation Contact Deepgram

Technology

Cisco operates some of the largest enterprise voice infrastructure in the world, from contact center platforms to collaboration and unified communications. Deepgram provides streaming speech-to-text for Cisco voice workloads where transcription accuracy and latency matter, with Nova-3 as the underlying model. Real-time speech recognition has to hold up in the conditions Cisco customers actually run: diverse accents, background noise, domain-specific vocabulary, and alphanumeric read-back where a single misheard digit breaks the call. Nova-3 is built for that production envelope, and Deepgram's voice-native foundational models are designed for the throughput a large contact center or voice-enabled product requires. For Cisco IVR platforms, Deepgram integrates through the open-source UniMRCP protocol, which lets enterprises drop Deepgram STT into existing Cisco voice flows without re-architecting the call path. The plugin and developer documentation are published below. If you are a Cisco customer or product team evaluating speech recognition, deepgram.com/contact-us is the direct route. Outlinks & Resources Deepgram Announces UniMRCP Integration to Power Modern Customer Experience Deepgram UniMRCP Plugin (developer docs) Deepgram UniMRCP server plugin (GitHub) Deepgram documentation Contact Deepgram

Alchemist Accelerator is the leading accelerator for enterprise B2B startups, with 650+ companies accelerated since 2012 and $5B+ raised across the portfolio. Alchemist's positioning sits directly on Deepgram's ICP: enterprise B2B, AI-SaaS, and deeptech founders building for enterprise buyers. For Alchemist portfolio companies building voice agents, contact center products, real-time meeting intelligence, or any voice-enabled enterprise feature, Deepgram provides the speech layer. Nova-3 for streaming STT, Aura-2 for TTS, and the Voice Agent API are available as a partner benefit, alongside Alchemist's other named partners (AWS, Google Cloud, Microsoft, Stripe, HubSpot, IBM). If you are an Alchemist founder, the Deepgram perk is accessible through your program portal. For founders beyond the accelerator's credit tier or for enterprise procurement terms, deepgram.com/contact-us is the direct route. Outlinks & Resources Alchemist Accelerator Alchemist Flagship Program Alchemist Portfolio Deepgram Startup Program Deepgram documentation Contact Deepgram

Secret is the SaaS deals platform for founders and startups, with 20,000+ members and negotiated discounts across 500+ tools. Founded in 2019, Secret runs a paid membership tier (lifetime or annual) that unlocks credits and discounts on the software stack a modern startup runs on. For Secret members building voice features into their products, Deepgram is on the deals catalog with $15,000 in credits over 12 months. Higher allocations are available based on usage. Nova-3 for streaming speech-to-text, Aura-2 for text-to-speech, and the Voice Agent API are all accessible against the credit balance. If you are a Secret member, the Deepgram deal is at joinsecret.com/deepgram. For founders not on Secret who want to evaluate Deepgram credits directly, the Deepgram Startup Program runs on its own terms at deepgram.com/startup-program. Outlinks & Resources Secret platform Deepgram on Secret Deepgram Startup Program Deepgram documentation Contact Deepgram

Technology

Axibator is an India-based student and early-stage founder community, with roughly 523 founders building across the country and a focus on Tier 2/3 cities and college campuses. The program runs an eight-week accelerator (Ignite) and on-ground residencies, with mentorship and community support for first-time builders. For Axibator founders building voice products (voice agents, transcription, multilingual assistants for the Indian market), Deepgram provides the speech layer. Nova-3 for streaming STT, Aura-2 for TTS, and the Voice Agent API are available as a partner benefit so student and early-stage founders can build voice features without standing up their own speech infrastructure. If you are an Axibator founder, claim the Deepgram benefit through the program's perks. For founders beyond Axibator who want to evaluate Deepgram on their own terms, the Deepgram Startup Program runs at deepgram.com/startup-program. Outlinks & Resources Axibator Axibator Ignite program Deepgram Startup Program Deepgram documentation Contact Deepgram

Development

Think41 is a full-stack generative AI consulting firm that builds secure, enterprise-ready AI agents. The team has partnered with Deepgram to power both their AI Agent Assist platform and their Voice AI Agent product, with Deepgram's STT and TTS providing the speech layer inside each. AI Agent Assist Think41's flagship Agent Assist system, powered by Deepgram, equips human agents with real-time, contextually aware guidance during live customer interactions and delivers post-call insights to streamline operations. The system includes: Real-time recommendations based on enterprise knowledge bases to assist agents during active calls Post-call analytics with summaries, sentiment scoring, and opportunity insights Direct integration into existing CRM and enterprise systems Voice AI Agent Beyond Agent Assist, Think41 has built an autonomous Voice AI Agent for fully human-like phone interactions. The product handles customer outreach, manages end-to-end flows, and knows when to hand off to a human, with Deepgram providing both the STT and TTS surfaces that make the conversation sound natural in real time. For enterprises evaluating an implementation partner Think41 brings senior engineers with deep generative AI experience and production deployment patterns refined across banking, retail, and enterprise tech customers. The Deepgram relationship means new model capabilities (Flux turn detection, Aura-2 voice options, multilingual coverage) reach client deployments quickly. If you are looking for an implementation partner to bring voice AI into production on Deepgram, Think41 is one of the firms we work with most directly. Contact details below. Outlinks & Resources Transforming Customer Interactions in Real Time: Think41 Partners with Deepgram to Power AI Agent Assist Think41 partner page Think41: Pioneering Full-Stack AI Solutions for Smarter Enterprise Operations (AI Apps) A Year of Voice AI at Scale (mentions Think41 deployments) Think41 Contact Deepgram

Technology

IBM selected Deepgram as its first fully embedded third-party speech model provider after customers asked for higher-accuracy STT and TTS, lower latency, and more flexible deployment than traditional cloud speech services could deliver. Deepgram's speech-to-text and text-to-speech models are now embedded into IBM watsonx, giving enterprises a modern voice AI stack on top of IBM's infrastructure. By embedding Deepgram inside watsonx Orchestrate Agent Builder, IBM clients can build voice agents and voice-enabled workflows on top of a real-time speech foundation that has been refined over more than a decade. Customers tap Deepgram's STT and TTS capabilities directly through their existing IBM environment, without adding extra vendors or commercial complexity. The integration covers contact-center conditions (accents, background noise, alphanumeric read-back) where speech accuracy makes the call usable, latency low enough to feel like a real conversation, and deployment across IBM Cloud, customer VPC, and self-hosted on regulated infrastructure. The integration is announced via joint press release and shipping inside watsonx Orchestrate today. If you are running IBM watsonx Orchestrate and evaluating voice, the integration is built in. Reach out via deepgram.com/contact-us for enterprise terms. Outlinks & Resources Deepgram and IBM Introduce Advanced Voice Capabilities for Enterprise AI Deepgram × IBM: Enterprise Voice AI Inside watsonx CX IBM watsonx Orchestrate Powered by Deepgram Contact Deepgram

Technology

Five9 is one of the largest cloud contact center platforms, and Deepgram provides the speech-to-text layer inside Five9 IVA Studio that powers agent assist, voicebot containment, and post-call analytics. In real-world tests, Deepgram is 2-4 times more accurate than alternative STT options inside Five9 for transcribing alphanumeric inputs (account numbers, policy IDs, tracking numbers, case IDs) that drive self-service interactions. After switching to Deepgram in Five9 IVA Studio, a major US healthcare provider doubled their user authentication rates because the speech recognition could accurately handle account-number read-back at scale. For Five9 customers, this translates into improved call resolution, higher self-service containment without escalating to a live agent, and meaningful cost savings. Adding Deepgram is a configuration step inside IVA Studio. Five9 is Powered by Deepgram. If you are running Five9 and evaluating speech recognition, the case study below has the technical detail. For benchmarks against your own audio, deepgram.com/contact-us is the direct path. Outlinks & Resources Five9 case study (deepgram.com/learn) Five9 case study (deepgram.com/customers) Five9 and Deepgram Partner to Revolutionize Contact Center AI Five9, Powered by Deepgram (partner page) Five9 IVA Studio Contact Deepgram