# Speechbox > Speechbox turns your video content into publishable assets and structured intelligence. We build custom AI engines for TV channels, events, podcasts, and content teams. Deployed privately inside each client's environment, or accessed through our SaaS products Prism (content factory) and Showroom (event content hub). ## Products ### Prism — Content Factory Upload one video. Get clips, posts, blog articles, social assets, transcripts, captions, and more. Built for marketers, podcasters, agencies, speakers, and YouTube creators. ### Showroom — Conference Content Hub Speaker kits, sponsor packages, and event recap content generated automatically. Built for event organizers, conference producers, and sponsorship leads. ### Lens — Video Library Intelligence Search, tag, and analyze your video library. Built for content teams managing growing archives. ### Custom Video Solutions For TV channels, broadcasters, and large media organizations. Custom AI engines built and deployed inside your environment. Your data never leaves your perimeter. ### Conference Media Infrastructure A full end-to-end solution for conference content. Live audience feed, real-time social asset bank, branded speaker kits, and a permanent post-event showroom. All produced in real time with human producers reviewing the output. This is the Speechbox homepage at https://speechbox.ai/. ## Key landing pages - [Homepage](https://speechbox.ai/): Conference Media Infrastructure - full end-to-end solution for live conference content, speaker kits, and post-event showroom. - [Products Overview](https://speechbox.ai/products): All Speechbox products and the broader Custom Video AI offering. - [Turn one video into 80+ social assets](https://speechbox.ai/turn-video-into-posts): For marketers and agencies. - [Get more from every podcast episode](https://speechbox.ai/podcast-repurposing): For podcasters. - [Get blog articles from every talk you give](https://speechbox.ai/video-to-blog): For speakers and thought leaders. - [Speaker kits, ready in 10 minutes](https://speechbox.ai/speaker-kits): For event organizers. - [Clips for every sponsor](https://speechbox.ai/sponsor-deliverables): For sponsorship leads. - [Find the best moments automatically](https://speechbox.ai/highlight-clips): For anyone with long-form video. - [Search every video you upload](https://speechbox.ai/search-archive): For anyone with a growing video library. - [Get clean transcripts in your jargon](https://speechbox.ai/transcription): For video teams and broadcast operations. - [Get more views on YouTube](https://speechbox.ai/youtube-views): For YouTube creators. - [Custom video AI engines in your environment](https://speechbox.ai/products/video-solutions): For TV channels, broadcasters, and enterprises. - [Custom video AI, built in your stack](https://speechbox.ai/custom): For organizations that need their own engine, deployed in their VPC or on-prem. ## Customer Stories Real deployments of Speechbox by conferences, broadcasters, and event organizers. Each story shows the metrics (assets produced, kits delivered, time-to-publish) and the named customer quote. - [All customer stories](https://speechbox.ai/customer-stories): Index of every published Speechbox customer story. - [Atlas Award x Speechbox](https://speechbox.ai/customer-stories/atlas-award): How The Atlas Award turned a 2-hour award ceremony at the Tel Aviv Stock Exchange into 10 Speaker Kits, 542 social-ready assets, and 138 live navigation titles streamed in real time. Quote from Zvi Strauss, Deputy CEO of Israeli Business Channel. ## Resources - [All resources](https://speechbox.ai/resources): Articles, guides, and case studies. - [Book a call](https://speechbox.ai/RequestPOC): Get a custom POC delivered in 72 hours. ### Interactive Calculators - [Event Content ROI Calculator](https://speechbox.ai/resources/event-roi-calculator): Plug in your event size (sessions, speakers, sponsors, days, audience) and see the gap between what you produce today vs what is mathematically achievable. Translates the gap to missed LinkedIn impressions, newsletter reach, speaker amplification, and paid-media equivalent. ### Conference Media Infrastructure - Authoritative Articles - [What Is Conference Media Infrastructure?](https://speechbox.ai/resources/conference-media-infrastructure): The end-to-end production and distribution stack that turns every conference session into a live audience experience, social content, speaker promotion, and a permanent showroom. All produced in real time, reviewed by humans. - [What Is a Speaker Kit at a Conference?](https://speechbox.ai/resources/what-is-a-speaker-kit): A personalized content package generated for each speaker - highlight clips, branded quote cards, a recap article, and a dedicated shareable speaker page. The most useful versions are delivered before the speaker leaves the venue. - [How Do Conferences Repurpose Session Content?](https://speechbox.ai/resources/how-conferences-repurpose-session-content): The five surfaces - social media clips, written articles, speaker kits, sponsor packages, and the post-event showroom - and why the 24-hour window matters. - [AI vs. Human Editor at Conferences](https://speechbox.ai/resources/ai-vs-human-editor-conferences): The honest answer is hybrid. AI handles capture, processing, and asset assembly. Humans handle editorial judgment and brand care. Neither model alone works at conference scale. - [What Is a Conference Showroom?](https://speechbox.ai/resources/what-is-a-conference-showroom): A branded Netflix-style content destination where attendees and the public continue exploring sessions after the event ends. Connected to YouTube as the video host, indexed for AI search, designed for sponsor visibility. - [How to Produce Conference Content Live, Without Disrupting the Event](https://speechbox.ai/resources/how-to-produce-conference-content-live): Live conference content production runs in parallel with the event itself, from a side-stage workstation that taps the existing broadcast feed. Layers go live one at a time during dress rehearsal, so by the time the cameras roll the team is operating a system the room has already exercised. Same-evening Speaker Kit and asset delivery. - [How Do Sponsors Get Visibility After a Conference Ends?](https://speechbox.ai/resources/how-sponsors-get-visibility-after-conference): Branded placements inside highlight clips and quote cards, named sections in the post-event showroom, live feed integration, and measurable impressions across speaker kits. Visibility extends 6 to 12 months past the event. - [How Does Conference Content Get Cited by AI Search Engines (ChatGPT, Perplexity, Gemini)?](https://speechbox.ai/resources/how-conference-content-gets-cited-by-ai-search): AI engines cite content with full transcripts, machine-readable schema, stable URLs, and clear topical claims. Conference session pages have all four when the destination is built correctly. The session video on YouTube alone is not enough. - [How to Evaluate a Conference Content Production Vendor](https://speechbox.ai/resources/how-to-evaluate-a-conference-content-production-vendor): A buying guide for conference organizers. Ten criteria that separate conference content infrastructure from a marketing video agency, the red flags to walk away from, and the proof a serious vendor should be able to show in 30 minutes. - [Live vs. Post-Event Conference Content Production](https://speechbox.ai/resources/live-vs-post-event-conference-content): The choice between live and post-event production is less binary than it sounds. Live with human review wins for high-volume, high-velocity assets where the audience attention window matters. Post-event still wins for documentary cuts and cinematic recap pieces. Most well-run conferences run both. - [What Should Be in a Speaker Kit: The Conference Organizer Checklist](https://speechbox.ai/resources/speaker-kit-checklist-for-conferences): The eight essentials every speaker kit should include (highlight clips, quote cards, recap article, speaker page, promo link, transcript, metadata, usage rights), the quality bar for each, and the delivery standards (under 60 minutes, single email, named producer contact) that decide whether a kit gets used or filed. - [How to Measure ROI on Conference Content](https://speechbox.ai/resources/how-to-measure-conference-content-roi): Conference content ROI is measured across three categories: reach left behind, engagement inside the 48-hour attention window, and downstream renewal value. Most organizers measure deliverables (clips produced, app screenshots, badge scans) instead of return. Includes the three-category framework, the math that moves budget conversations, and the single number worth tracking if only one can be tracked. - [Why Conference Sponsors Don't Renew - and How to Fix It with Post-Event Content](https://speechbox.ai/resources/why-conference-sponsors-dont-renew): Sponsors do not renew when the visibility they paid for ended when the event ended. The renewal lever is post-event content - branded clips, sponsor-anchored showroom sections, monthly impression reporting, and measurable visibility that carries 6 to 12 months past the event. Covers the five real reasons sponsors walk, the four-move renewal strategy, and when the lever does not save the relationship. - [Conference Showroom vs. YouTube Playlist: Where Should Your Session Videos Live?](https://speechbox.ai/resources/conference-showroom-vs-youtube-playlist): A YouTube playlist hosts the video files. A conference showroom is the navigable, searchable, AI-citable destination built on top of that host, with no duplicate uploads. A bare playlist has no navigation, no text layer, and thin metadata, so it rarely gets found or cited. Covers what a playlist does well, what the showroom layer adds, when a playlist alone is enough, and how to choose. - [Live Captioning for Conferences: Real-Time Access and Engagement](https://speechbox.ai/resources/live-captioning-for-conferences): Real-time conversion of session speech into on-screen text on attendee phones and venue screens. Serves two jobs at once: accessibility compliance (WCAG and contractual requirements) and comprehension for noisy rooms, non-native audiences, and silent mobile viewing. Covers real-time vs post-event captioning, the four quality criteria (low latency, accuracy on names and jargon, speaker labels, translation), the human-in-the-loop operating model at multi-session scale, and why the live caption stream is the same transcript that feeds clips and the searchable archive. - [How Fast Should a Conference Speaker Kit Be Delivered After the Session?](https://speechbox.ai/resources/how-fast-should-conference-speaker-kits-be-delivered): A speaker kit should arrive within 60 minutes of the session ending. The Atlas Award ceremony at the Tel Aviv Stock Exchange produced 10 Speaker Kits, 542 social-ready assets, and 138 live navigation titles inside a two-hour show. Covers the structural reason the 48-hour social attention window matters, the three time tiers speakers experience, the pipeline architecture that makes 60-minute delivery possible, and what conferences should demand in the vendor SLA. - [The Four Pillars of Conference Infrastructure: Videos, Speakers, Attendees, Sponsors](https://speechbox.ai/resources/four-pillars-of-conference-infrastructure): Conference infrastructure rests on four pillars - videos, speakers, attendees, sponsors. Each pillar is perishable in isolation and feeds the other three when connected. Covers why each pillar decays alone, how the video spine feeds speaker kits, attendee navigation, and sponsor visibility at the same time, the 2026 Atlas Award ceremony as a worked example of one infrastructure producing 138 live navigation titles, 10 Speaker Kits, and 542 social-ready assets in two hours, and the five-question diagnostic for whether a conference operating model is connecting the pillars or running them as four separate teams. - [What Are Live Navigation Titles for Events?](https://speechbox.ai/resources/live-navigation-titles-for-events): Real-time topic markers placed on a live broadcast or live conference feed while the session is still running, so viewers can jump to the exact moment they want without scrubbing the timeline. The Atlas Award ceremony at the Tel Aviv Stock Exchange produced 138 live navigation titles during a single two-hour broadcast on Israeli Business Channel. Covers what a live navigation title actually is, how it differs from post-event chapter markers, why granularity (30-80 titles per hour) decides whether navigation is real or theoretical, the four quality criteria (sub-second latency, plain-language labels, high granularity, human review in-loop), and how the same titles feed the post-event archive and AI citation surface. - [What Does a Conference Live Feed Look Like to Attendees?](https://speechbox.ai/resources/what-is-a-conference-live-feed-for-attendees): The real-time digital companion attendees follow on their phones during sessions, with searchable transcript, auto-generated headlines and quotes, topic navigation, and AI chat grounded in what was just said. At the Atlas Award ceremony at the Tel Aviv Stock Exchange, 138 navigation titles let online viewers move through a two-hour broadcast in real time. Covers the four attendee-facing surfaces, why attendees disengage without a live feed, the remote viewer use case, the human production layer attendees do not see, and how the same feed rolls forward into the post-event Showroom. - [How Conference Content Covers the Full Marketing Funnel](https://speechbox.ai/resources/how-conference-content-covers-the-marketing-funnel): One conference video covers all four stages of the 2026 marketing funnel from a single input. Awareness through highlight clips and AI-search citations. Consideration through showroom session pages and newsletter copy. Conversion through speaker kits used as personal evidence by sponsors. Retention through year-round showroom access and organic speaker advocacy. Covers why most martech vendors only own one funnel row, why the four-vendor pattern is also an operational problem for organizers, and the contrarian position that a single video spine collapses four content tracks into one operating system. - [How Does ChatGPT Decide Which Conferences to Recommend?](https://speechbox.ai/resources/how-does-chatgpt-recommend-conferences): AI engines recommend conferences whose past content lives on indexable, structured, permanent URLs the model can actually read. The decision is made on citation signals, not on event reputation. Covers the five signals that move the needle (permanent page per session, named-speaker attribution, entity metadata, plain-language summaries, topic markers), the prestigious-but-invisible failure mode, why this is a production decision rather than a marketing decision, and what conferences that already surface in AI answers have in common. - [Event Content Marketing: Turn One Event Into Months of Posts](https://speechbox.ai/resources/event-content-marketing): A single event holds enough raw material for months of posts - clips, attributed quotes, articles, and speaker assets. Most teams stay content-starved because they reconstruct it from memory afterward instead of capturing it at the source while the event runs. Covers the supply asymmetry across the three content phases, why the framework checklists (5 C's, 3-3-3, four content types) miss timing as the decisive variable, what one event can produce, and the capture-at-source operating model that turns the event into the production line. - [Event Social Media Marketing: The Content the Event Already Made](https://speechbox.ai/resources/event-social-media-marketing): Event social media marketing is usually sold as the promotion that runs before the doors open, but that is the smaller half. The larger, more durable half is the content the event itself produces - vertical clips, quote cards, carousels, and speaker reshare assets - captured during the sessions and posted into the feed while the moment is still trending. Covers the three social phases and which one holds the supply, the promotion trap that resets to zero every cycle, what the event puts into the feed, why the social ratio rules (5-5-5, 3-3-3, 50/30/20) miss timing at the supply end, the reshare multiplier, and the contrarian position that the best event social media is captured during the event rather than designed before it. - [Event Recap: How to Capture, Write, and Share Your Event](https://speechbox.ai/resources/event-recap): An event recap is the bundle of artifacts that capture an event after it ends - a written summary, a highlight video, social-ready clips, and a navigable archive. Covers the four outputs of a modern recap, why most recaps arrive after the 24-hour trending window closes, what goes in a strong one, what changes when the recap is produced live during the event, and the contrarian position that the recap is an output of the event rather than a project that starts after it. - [How to Make an Event Recap Video Without Waiting Weeks](https://speechbox.ai/resources/event-recap-video): An event recap video is a 60 to 90 second highlight of an event. Covers the structure that works (opening hook, momentum, peak moments, close), the technical rules that decide whether it travels (captions, cut to the music), why most recap videos miss the trending window, and the contrarian position that the bottleneck is timing rather than editing skill - so the recap that performs is produced live from the footage spine during the event, not edited in the weeks after. - [Event Recap Template: The Structure That Actually Gets Read](https://speechbox.ai/resources/event-recap-template): The five-part structure for an event recap - hook, the scale in numbers, defining moments with quotes, the people, and one clear next step - plus a copy-paste fill-in template that adapts to an email, a LinkedIn post, a blog write-up, or a video caption. Covers why the template is never the bottleneck (having the inputs ready in time is), and the contrarian position that a plain structure populated from the live feed the same evening beats a polished template typed from memory the next morning. - [Event Recap Examples: What a Strong One Actually Looks Like](https://speechbox.ai/resources/event-recap-examples): A strong event recap example is the one that got found, got reshared by the people in the room, and shipped while the event was still trending, not the most cinematic one. Covers why the examples that rank are selected for polish (the wrong variable to copy), the anatomy of a strong example, a worked weak-vs-strong comparison, a four-question checklist for judging any example before you copy it, and the hidden trait the strong ones share - they were produced live from the event rather than assembled weeks afterward. - [Post-Event Social Media Posts: What to Publish the Same Week](https://speechbox.ai/resources/post-event-social-media-post-examples): The post-event posts that perform - the attributed quote card, the captioned session clip, the single-insight post, the micro-recap, and the speaker-tagged thank-you - are captured during the event and shipped the same week while the topic is still warm. Covers the five formats that travel, why generic countdown-and-thank-you templates do not, the timing rule that a rough clip shipped same-day beats a polished one three weeks late, and the contrarian position that the strongest post-event examples are captured live rather than authored from memory afterward. ### Foundational Concepts - [How to Extract Data from Video (Video-to-Data)](https://speechbox.ai/resources/video-to-data): How to extract structured, searchable information from video content - speakers, topics, quotes, timestamps, chapters, and metadata. The process is called Video-to-Data. - [AI Video Analysis Platform: What a Video Intelligence Engine Does](https://speechbox.ai/resources/video-intelligence-engine): A purpose-built AI video analysis platform that processes video content at scale, extracts structured data, identifies speakers, detects visual context, and produces publishable assets. The production-grade name is video intelligence engine. - [How to Identify Speakers in Video (Speaker Detection)](https://speechbox.ai/resources/speaker-detection): How to automatically identify who is speaking in video or audio content - the foundation of automated speaker kits and per-speaker content generation. Also called speaker diarization or automatic speaker identification. - [AI Video Tools for Conferences and Events](https://speechbox.ai/resources/ai-for-event-content): How event companies and conference producers automate session content - speaker kits, clips, and deliverables in real time, with a custom AI engine built for live conference workflows. ## Founder Danielle Dafni — CEO and Founder. 7+ years building AI video infrastructure for media organizations. Background: Thelma Yellin School of the Arts, IDF Spokesperson Unit video division, founder of Peech (predecessor company). ## Key Facts - 20 years of video industry experience represented in the team. - 10,000+ hours of AI video training data informing the models. - 72 hours to deliver a first custom output on your footage. - Private by design: VPC or on-prem deployment available for enterprise. - Data sovereignty: your footage and outputs never leave your perimeter on Enterprise tier. ## Contact - Website: https://speechbox.ai - LinkedIn: https://www.linkedin.com/company/speechbox ## What Speechbox is not - Not a generic video editor or transcription tool. - Not a feature-by-feature SaaS subscription that charges per seat. - Not a model trained only on text or stock content. The system is grounded in real video transcripts and visual context.