# Everlit - Complete Product Information

> Distribution and engagement platform for the written word, built for newsrooms, publishers, and brands.

Website: https://everlit.audio
Contact: hello@everlit.audio
Studio: https://studio.everlit.audio

---

## Company Overview

Everlit is a B2B SaaS distribution and engagement platform for the written word. It turns written content into gorgeous audio, social video, and podcasts and distributes it everywhere an audience already listens—podcast apps, social channels, and the web—driving 2-3x longer engagement. The platform serves newsrooms, publishers, universities, school districts, brands, and independent creators who want to reach audiences across audio, social, and podcast without manual production work.

## Core Product: Text-to-Audio

Everlit's primary function is converting written articles, blog posts, newsletters, and other text content into professional-quality audio. The process is fully automated ("zero-touch workflow") — publish your article and Everlit handles the rest.

### AI Voices & Voice Cloning
- Library of natural-sounding AI voices
- Voice cloning technology to create custom voices matching your brand
- Support for 70+ languages and regional accents
- Professional audio polishing (noise reduction, leveling, EQ)

### Embedded Audio Player
- Customizable, responsive audio player widget
- Embeds directly into any website or CMS
- Configurable appearance: standard and mini modes
- Custom title intros, icons, and branding options
- Click-to-play functionality

### Smart Playlists
- Automatically generated playlists from your content
- CarPlay compatible for in-car listening
- Dynamic updates as new content is published
- Mobile-optimized playback

### Custom Intro/Outro
- Add branded intro and outro music to every audio article
- Professional transitions between content and ads

## Everlit MCP Server (Agentic Interface) & TTS API

Everlit offers two developer products built on the same Everlit Voice AI model. The MCP (Model Context Protocol) server is an agentic interface to a whole Everlit account: Claude, Cursor, Claude Code, VS Code or any MCP client connects with OAuth and runs publishing workflows in plain language—publish audio articles from a URL or text (page extraction, narration in the publication's configured voice with byline, disclaimer, and intro/outro music) to a public player, edit or take down any Studio article, change publication defaults, curate playlists, fill podcast feed metadata, switch distribution on and track directory submissions, point a publication at an RSS/Atom feed so every new item is narrated automatically, maintain a pronunciation library, manage music beds and voices (including cloning), create speech, and read usage. The TTS API is the voice model on its own: studio-quality text-to-speech through a simple REST API for your own products. Product page: https://everlit.audio/api. Reference docs: https://everlit.audio/docs/tts-api (Markdown: https://everlit.audio/docs/tts-api.md, OpenAPI 3: https://everlit.audio/docs/tts-openapi.yml) and https://everlit.audio/docs/articles-api (Markdown: https://everlit.audio/docs/articles-api.md).

### Pricing
- $8 per million characters for Latin, Cyrillic, and Greek scripts, pay-as-you-go from a prepaid balance. No monthly fee, no plan required, credit never expires.
- Script-weighted billing keeps every language at about the same price per finished hour of audio: Arabic, Hebrew, and Indic scripts ×1.5 ($12/M), Korean and Thai ×2 ($16/M), Japanese ×2.5 ($20/M), Chinese and Cantonese ×3 ($24/M). Counted with standard Unicode script ranges, so bills are reproducible from your own request logs; weights only ever change in the customer's favor.
- Examples (English): a 1,200-word article ≈ $0.04; a 10,000-word long read ≈ $0.34; a 90,000-word audiobook ≈ $3.09; 1,000 articles a month ≈ $42.
- Top up from $5 to $10,000 via Stripe; optional automatic top-up and a low-balance warning email. Failed jobs are never charged; word-level timestamps, webhooks, and cloned voices cost nothing extra.
- Operational limits (safety rails, raised per account on request to support@everlit.audio): 60 speech jobs created per minute, 300 reads per minute, 4 concurrent jobs, 20,000 characters per request.

### Capabilities
- REST API at https://api.everlit.audio: POST /v1/speech creates a job (202; format mp3/wav/opus, speed 0.5–2.0, paragraph_pause 0–5 s, per-request pronunciations, inline markup [everlit-pause:N], [everlit-mixable:N], [everlit-mixable-background:N], [everlit-subvoice-open:N]…[everlit-subvoice-close]), POST /v1/speech/estimate prices a request and predicts duration without creating a job, GET /v1/speech/{id} polls it (and mints a key-free 24 h download_url), GET /v1/speech/{id}/audio downloads the file, GET /v1/speech/{id}/timings returns word-level timestamps, GET /v1/speech/{id}/captions returns SRT, VTT or JSON cues, POST /v1/audio/speech is a synchronous OpenAI-compatible endpoint (point the OpenAI SDK at base_url https://api.everlit.audio/v1), plus GET/POST/DELETE /v1/voices (cloning with consent), GET /v1/mixables, GET/PUT/DELETE /v1/pronunciations, GET /v1/languages, GET /v1/usage, GET /v1/account, and DELETE /v1/speech/{id}.
- Authentication: Authorization: Bearer api_… keys minted in Everlit Studio (API → Keys), shown once and stored hashed.
- 69 languages, including every official EU language, plus Arabic, Hebrew, Persian, Urdu, Hindi, Bengali, Tamil, Telugu, Japanese, Korean, Chinese, Cantonese, Thai, Vietnamese, Indonesian, Swahili, and more. The live catalog with billing weights is at GET /v1/languages.
- Word-level timestamps in 25 languages (Bulgarian, Croatian, Czech, Danish, Dutch, English, Estonian, Finnish, French, German, Greek, Hungarian, Italian, Latvian, Lithuanian, Maltese, Polish, Portuguese, Romanian, Russian, Slovak, Slovenian, Spanish, Swedish, Ukrainian). Words come from the submitted text verbatim, not from a speech recognizer, so captions and karaoke highlighting match the script exactly. Coverage expands over time.
- Voice cloning: POST /v1/voices with a 15 s+ single-speaker sample (upload or sample_url), the subject's name, relationship and a consent attestation; the clone is private to the account, appears in the catalog (owned_by: publisher), bills at the same rate, and can be removed by Everlit later with a client-visible reason. Default cap 10 clones per account.
- Billing modes: every paid Studio plan (Independent, Network, Media, Enterprise) includes speech with no per-character charge and no cap (billing.mode "plan"; concurrency = number of publications). Accounts without a plan pay as you go.
- Signed webhooks (callback_url, HMAC-SHA256 X-Everlit-Signature, retried for ~30 minutes), Idempotency-Key support (30 days), RateLimit-* headers on every response, and audio URLs that are known at submit time and stable for 30 days.
- Asynchronous by design: jobs typically finish in 15–60 seconds after a full render and automatic quality verification. Ideal for articles, archives, audiobooks, e-learning, and publish-time pipelines; not for realtime conversational agents or IVR.
- No SDK required—copy-paste examples in Python, Node, Ruby, and curl.

### MCP server
- Everlit gives AI tools and workflows the power of a full audio operation; the MCP server is the interface to it, standalone and not tied to the TTS API. It exposes 52 tools across articles, publications, playlists, podcast feeds and directories, RSS ingestion, pronunciations, music beds, speech, voices and account, so an agent can run an entire audio publishing workflow from one conversation.
- Endpoint: https://api.everlit.audio/v1/mcp (Streamable HTTP, stateless). Authenticate with OAuth 2.1 (authorization code + PKCE, dynamic client registration; you log in to Everlit Studio and approve the connection — no key to copy) or with the same Bearer API key the REST API takes. Without credentials the server still answers tools/list with two tools, request_access and get_access_request, so an agent can sign its user up from inside the conversation; initialize without credentials returns 401 with a WWW-Authenticate challenge that starts the OAuth login.
- get_account (also GET /v1/account) is the first call to make: publisher, key, entitlements (whether TTS is enabled), billing mode, balance and price, enforced limits (max characters per request, concurrency, rates), plan quotas, and links to fix any of it. Errors carry next_step and resolution_url so an agent can hand the user a link.
- Works with Claude.ai (Settings → Connectors → Add custom connector), Claude Desktop, Claude Code, Cursor, VS Code, and any MCP-capable client. Claude Code install with OAuth: claude mcp add --transport http everlit https://api.everlit.audio/v1/mcp, then /mcp → everlit → Authenticate. Static key: add --header "Authorization: Bearer api_YOUR_KEY".
- 52 tools. Articles: create_article, get_article, list_articles, update_article, delete_article (covers every Studio article: widget, WordPress, RSS, uploads, API; article text accepts inline [everlit-pause:N], [everlit-mixable:N], [everlit-mixable-background:N] and [everlit-subvoice-open:N]…[everlit-subvoice-close] markup). Publications: list_publications, get_publication, update_publication. Playlists: list_playlists, get_playlist, create_playlist, add_to_playlist, remove_from_playlist, move_in_playlist, update_playlist, delete_playlist. Podcast feeds: get_feed, update_feed, activate_feed, deactivate_feed, list_feed_platforms (status per directory: Apple Podcasts, Spotify, Pocket Casts, Amazon Music, YouTube, Overcast, each with a submission_url to hand a person), mark_feed_submitted, mark_feed_live, reset_feed_platform, ping_feed (Overcast re-crawl). RSS ingestion: list_ingestion_feeds, create_ingestion_feed (publication + RSS/Atom url, poll_interval_minutes 5–1440, max_items, oldest_date; every new item is narrated with the publication's defaults and counts against the article quota), get_ingestion_feed, update_ingestion_feed, delete_ingestion_feed, poll_ingestion_feed, list_ingestion_items. Pronunciations: list_pronunciations, set_pronunciation, delete_pronunciation. Music beds: list_mixables, create_mixable, delete_mixable. Speech: create_speech, estimate_speech, get_speech, list_speech, get_speech_timings, get_speech_captions, delete_speech. Voices: list_voices, create_voice, get_voice, delete_voice. Account: get_account, get_usage, list_languages. Without credentials: request_access, get_access_request.
- Resources: everlit://docs/tts-api, everlit://docs/articles-api, everlit://languages. Prompts: narrate_text(text, voice?), publish_from_url(url, publication?). Tools carry readOnlyHint / destructiveHint annotations so clients confirm deletes and never confirm lists.
- Example workflows: "Publish these five URLs to The Daily, pin them to Morning Briefing in order, give me the embed for the lead story." "Make a playlist of everything tagged climate, fill it automatically, set the show metadata, activate the podcast feed, and give me the Apple and Spotify submission links." "Narrate everything our partner's RSS feed publishes from now on into the Wire publication." "Add these surnames to the pronunciation library, switch Sports to Andrew's voice, report this month's speech spend."
- Running jobs return poll_after_seconds so agents wait politely instead of spinning; tool errors carry the same structured error envelope as REST plus http_status and retry_after. Tool calls are metered, limited, and billed exactly like REST calls.

### Getting started
Demand for the API is high, so Everlit reviews every access request (usually within a business day) to keep capacity where customers expect it.
1. Request an API key at https://everlit.audio/api#request-key (use case and expected monthly volume). Most requests hear back within a business day. Agents can do this without a key: POST https://api.everlit.audio/v1/access-requests with email, use_case and product, then poll GET https://api.everlit.audio/v1/access-requests/{id} (status pending / approved / denied, with next_step).
2. Once approved, sign in to Everlit Studio (https://studio.everlit.audio), open Billing → TTS API, add a card and starting credit (from $5), then create an api_ key under API → Keys.
3. POST to /v1/speech or connect the MCP server. Existing Studio customers can skip the key entirely for MCP: connect with OAuth and log in. Usage and generations appear in Studio within minutes.
Questions: support@everlit.audio.

## Distribution

### Podcast Distribution
Everlit automatically syndicates your audio content to all major podcast platforms:
- Apple Podcasts
- Spotify
- Pocket Casts
- Any RSS 2.0 compatible platform
- Full Podcast 2.0 support
- Proper ID3 tags and metadata

### Social Video Generation
Everlit transforms audio content into optimized video for social media:
- Horizontal format (16:9) for YouTube
- Square format (1:1) for Instagram and Facebook
- Vertical format (9:16) for TikTok, Instagram Reels, YouTube Shorts
- Automatic generation from published audio

## Monetization

### Ad Types
- Programmatic ads via Google Ad Manager (GAM) integration
- Direct-sold sponsorships
- House ads for self-promotion

### Ad Placements
- Preroll (before content)
- Midroll (during content)
- Postroll (after content)
- After title placement
- After byline placement
- Custom location placements

### Campaign Management
- Run multiple campaigns simultaneously
- Run-of-site (ROS) targeting
- IAB category targeting
- Keyword-based targeting and exclusions
- Companion banners
- Hero image logo overlays

### Analytics
- Comprehensive dashboards
- Listen counts and engagement metrics
- Export to favorite BI systems
- Real-time reporting

## CMS Integrations

Everlit works with any content management system through an effortless, zero-touch workflow:
- WordPress (native plugin)
- Every other CMS — including Arc Publishing, Naviga, Brightspot, Contentful, and HubSpot — uses Everlit's zero-touch JavaScript embed: a tiny code block you paste in once, after which audio is generated automatically every time you publish. No plugin required.
- Custom platforms can also connect via the Everlit MCP server and TTS API (https://everlit.audio/api).

## Customers

Everlit is trusted by organizations including:
- The Texas Tribune
- Hearst Newspapers
- San Francisco Chronicle
- SFGate
- CT Insider
- Chron
- Advance Local
- The Narwhal
- Auburn University
- Content Marketing Institute
- Lookout Local
- Oklahoma Watch
- Babylist
- Pflugerville School District
- Hiscox
- Shaw Media
- Religion News Service
- Ora de Sibiu (Romanian-language newspaper)
- Newspack

## Target Markets

1. **Local & National News** — Newsrooms of all sizes converting articles to audio
2. **Magazines & Publishers** — Print and digital publications adding audio channels
3. **Universities & Colleges** — Campus communications, research papers, newsletters
4. **School Districts (K-12)** — Making educational content accessible via audio
5. **Brands & Marketers** — Content marketing teams amplifying written content
6. **Independent Creators** — Bloggers, writers, and solo publishers

## Engagement Statistics
- Audio content drives 2-3x longer engagement compared to text alone
- Higher retention rates for audio consumers
- 24/7 accessible content for on-the-go audiences

## Technical Details
- Web-based platform (no software installation required)
- MCP server (agentic interface, 52 tools) so AI agents can publish articles, curate playlists, run podcast feeds and directory submissions, auto-narrate RSS feeds, and manage pronunciations, voices and music
- TTS REST API for custom integrations (included on paid plans; otherwise $8 per million characters, pay-as-you-go)
- Articles REST API to publish complete audio articles from a URL or text (counts against the publication's monthly plan)
- RSS 2.0 and Podcast 2.0 compliant feeds
- ID3 tag support for proper metadata
- Responsive player works on all devices and screen sizes
- Audio players and playlists are fully accessible and meet WCAG 2.2 AA standards (keyboard navigation and screen-reader support)
- Rich player can be embedded anywhere, including social, and stories are pushed out to every channel automatically on publish

## Frequently Asked Questions

Full FAQ: https://everlit.audio/faq

Q: What is Everlit?
A: Everlit is a distribution and engagement platform for the written word. It turns articles, blog posts, and newsletters into audio, social video, and podcasts, then distributes them everywhere your audience already listens—your website, podcast apps, and social channels.

Q: What is text-to-audio, and how does Everlit use it?
A: Text-to-audio converts written content into natural-sounding spoken audio. Everlit does this automatically the moment you publish, using high-quality AI voices with professional polishing, then distributes the result to your on-site player, podcast apps, and social channels with no manual production.

Q: How is Everlit different from basic text-to-speech?
A: Basic text-to-speech just reads words aloud; Everlit turns content into a distribution channel—adding broadcast-quality AI voices and voice cloning, an embedded player, smart playlists, automatic podcast syndication, social video, analytics, and ad monetization in one zero-touch workflow.

Q: Is AI-generated audio good enough for professional news and journalism?
A: Yes. Everlit uses high-quality AI voices with professional polishing (leveling, noise reduction, natural pacing) so every article sounds broadcast-ready. Newsrooms including The Texas Tribune, the San Francisco Chronicle, and Hearst Newspapers publish their journalism as Everlit audio.

Q: Can I choose or customize the voice, and can I upload my own narration?
A: Yes. Choose from high-quality AI voices, clone a custom voice to match your brand or staff narrators, use conversational podcast mode, or upload your own recorded narration.

Q: How many languages does Everlit support?
A: More than 70 languages plus regional accents.

Q: Can I turn my blog or newsletter into a podcast automatically?
A: Yes. Everlit automatically turns blog posts and newsletters into podcast episodes and syndicates them to Apple Podcasts, Spotify, Pocket Casts, and any RSS-based app—with artwork, metadata, and a proper podcast RSS feed—every time you publish.

Q: How do I publish my articles to Spotify and Apple Podcasts?
A: Everlit does it automatically from a single publish, using RSS 2.0, Podcast 2.0, and ID3 standards. There is no manual uploading or feed management.

Q: Can I add a "listen to this article" button to my website?
A: Yes. Everlit provides an embeddable, brand-matched audio player that installs with a small JavaScript snippet or the native WordPress plugin and appears on new articles automatically.

Q: What CMS does Everlit work with, and is there an API?
A: Every CMS. WordPress gets a native plugin; every other platform (Arc, Naviga, Brightspot, Contentful, HubSpot, custom) uses a zero-touch JavaScript embed. An MCP server (for AI agents) and a TTS REST API (for custom code) are also available.

Q: Does Everlit offer a text-to-speech API?
A: Yes. The Everlit TTS API returns studio-quality MP3 narration from a simple REST endpoint, in 69 languages with cloned voices and free word-level timestamps, at $8 per million characters pay-as-you-go. It is asynchronous (15–60 s per job), built for articles, archives, audiobooks, and pipelines rather than realtime conversation. See https://everlit.audio/api.

Q: Does Everlit have an MCP server for AI agents?
A: Yes. https://api.everlit.audio/v1/mcp is a Streamable HTTP MCP server (OAuth login or a static API key) and a standalone agentic interface to an Everlit account, with 52 tools: articles (create, get, list, update, delete—any Studio article), publications (list, get, update defaults), playlists (list, get, create, add, remove, move, update, delete), podcast feeds (get, update metadata, activate, deactivate, list directory status, record submitted/live, reset, ping Overcast), RSS ingestion (list, create, get, update, delete, poll now, list items), pronunciations (list, set, delete), music beds (list, create, delete), speech (create, estimate, get, list, timings, captions, delete), voices (list, create/clone, get, delete) and account (get_account, get_usage, list_languages); plus request_access and get_access_request for a client with no credentials. It also serves resources (both API guides, the language catalog) and prompts (narrate_text, publish_from_url). Claude, Cursor, Claude Code, VS Code and other MCP clients can run whole workflows from one message: publish a set of URLs, pin them to a playlist, switch on the podcast feed, fix pronunciations, change a publication's voice, and report usage.

Q: How much does the Everlit TTS API cost?
A: $8 per million characters for Latin, Cyrillic, and Greek scripts, with other scripts weighted (×1.5 to ×3) so every language costs about the same per hour of audio. Prepaid pay-as-you-go from $5, no monthly fee, credit never expires, failed jobs never charged.

Q: Can I add audio to articles I've already published?
A: Yes. New articles convert automatically, and the Everlit team can help generate audio for your existing archive during onboarding.

Q: How do publishers make money with audio on Everlit?
A: Three ways—programmatic ads through Google Ad Manager, direct-sold sponsorships, and house ads—with preroll, midroll, postroll, custom placements, and companion banners.

Q: What analytics does Everlit provide?
A: Listen counts, completion rates, device breakdowns, and audience insights, with dashboards you can export to your business intelligence tools.

Q: Does audio actually increase engagement and revenue?
A: Yes. Audio drives 2-3x longer sessions and roughly 40% higher retention than text alone; at The Texas Tribune, 84% of listeners finish full episodes.

Q: Will adding an audio version hurt my SEO or create duplicate content?
A: No. The written article stays the canonical source; the embedded player increases time-on-page and engagement (signals that support SEO), and audio opens new discovery surfaces via podcast apps and social video.

Q: Does adding audio help with ADA and accessibility compliance?
A: Yes. Audio makes content available to people who are blind, low-vision, or dyslexic, and Everlit's players meet WCAG 2.2 AA standards with keyboard navigation and screen-reader support.

Q: Is Everlit secure and privacy-compliant?
A: Everlit follows enterprise-grade security practices and offers GDPR compliance, a Data Processing Agreement, and a published subprocessor list. See https://everlit.audio/legal.

Q: Is Everlit's voice cloning consent-based?
A: Yes. Voice cloning is for brands and on-staff narrators who opt in, to create custom voices you have the rights to use—not to impersonate anyone without permission.

Q: How much does Everlit cost, and how do I get started?
A: Everlit offers tiers based on the features an organization needs, from independent creators to large newsrooms and brands. Book a demo via the contact form or hello@everlit.audio.

## Documentation

Developer and integration docs: https://everlit.audio/docs

- TTS API reference (REST + MCP server, billing, rate limits, webhooks): https://everlit.audio/docs/tts-api — Markdown: https://everlit.audio/docs/tts-api.md — OpenAPI: https://everlit.audio/docs/tts-openapi.yml
- Articles API reference (publish full audio articles from a URL or text, REST + MCP tools, quotas, webhooks): https://everlit.audio/docs/articles-api — Markdown: https://everlit.audio/docs/articles-api.md
- Auto Audio (one-script listen-to-this-article player): https://everlit.audio/docs/auto-audio
- Web integration (embeds, players, CMS): https://everlit.audio/docs/web-integration
- Mobile integration (iOS/Android webview): https://everlit.audio/docs/mobile-webview
- On-demand playlists: https://everlit.audio/docs/on-demand-playlists
- Podcast distribution (Apple, Spotify, Amazon): https://everlit.audio/docs/podcast-distribution
- Voice library: https://everlit.audio/docs/voices
- Advertising (campaigns, ads, creatives, contexts): https://everlit.audio/docs/advertising

## Dispatches (Blog)

Everlit publishes in-depth articles at https://everlit.audio/dispatches covering:

- How The Texas Tribune turns its journalism into audio with Everlit (customer story: 84% completion, Newspack publishing, new revenue)
- Ad Auris acquisition and migration path to Everlit
- Audio monetization with programmatic ads, sponsorships, and GAM integration
- CMS compatibility (WordPress, Arc, Naviga, Brightspot, Contentful, HubSpot)
- WordPress plugin deep dive and Newspack partnership
- Social video generation for TikTok, Instagram Reels, YouTube Shorts
- Podcast distribution to Apple Podcasts, Spotify, and RSS platforms
- AI voices, voice cloning, and 70+ language support
- Analytics, engagement metrics, and audio content strategy
- Audio publishing for Spanish-language publishers (US Hispanic, Latin America, Spain)
- Audio publishing for European multilingual newsrooms

### Multilingual Dispatches

All 11 Dispatches articles are available in 7 languages with localized content and SEO-optimized slugs:

- English: https://everlit.audio/dispatches
- Spanish: https://everlit.audio/dispatches/es
- French: https://everlit.audio/dispatches/fr
- German: https://everlit.audio/dispatches/de
- Italian: https://everlit.audio/dispatches/it
- Portuguese: https://everlit.audio/dispatches/pt
- Dutch: https://everlit.audio/dispatches/nl

Each language version is written natively for its target market (e.g., German articles reference the DACH market, Portuguese articles address both Brazilian and European audiences). Articles use hreflang tags to link all language versions together for multilingual SEO.

## Contact
- Email: hello@everlit.audio
- Website: https://everlit.audio
- Studio login: https://studio.everlit.audio
- API support: support@everlit.audio
- API status: https://status.everlit.audio