Turn anything into audio — then refine it.

• Quick Read: paste text → generate audio quickly (single job).
• Studio: multi-segment timeline editor for production workflows.
• Feeds: RSS/Atom + crawl-based “inbox” for articles/podcasts/newsletters (with optional YouTube channel support).
• Library: browse/organize voices, favorites, collections, and presets (synced).
• Projects: manage saved work and play/export long-form outputs.

• A single persistent player used across the app for:
  • Quick Read audio,
  • Studio segment playback,
  • merged project playback,
  • feed-item audio playback,
  • YouTube playback mode (separate from TTS audio).

• Voice selection (simple + advanced voice modes) with styles, rate, pitch, and volume controls.
• Optional SSML input mode.
• Real-time generation progress with job resumption after refresh.
• Word-level highlighting view when timing data is available.
• Global persistent playback (doesn’t stop when navigating to other pages).

• Segment-based timeline (text segments + pause segments).
• Per-segment voice settings: voice, style, rate, pitch, volume, SSML vs plain text.
• Batch operations: reorder, select many, bulk delete, collapse/expand, “apply settings to all”.
• “Unified Text View”: edit all segment text in one continuous editor with non-editable segment markers.
• Focus mode for distraction-free editing.
• Segment-level generation: generate, regenerate, cancel, play-while-generating.
• Project-level Word Sync toggle (controls whether word-level timing is generated/used).
• Segment waveform timeline (“segmented waveform”) for fast navigation/seek to a specific segment/time.
• Per-segment visual SSML editor (build SSML with UI for breaks, emphasis, prosody, say-as, phonemes, substitutions) + manual SSML editing.

• Import formats:
  • Timeline JSON (.json) project imports/exports.
  • Text/Markdown (.txt, .md) with formatting preservation for emphasis.
  • PDF (.pdf) via client-side extraction (PDF.js).
  • Word (.docx) with formatting/headings (mammoth.js → Markdown).
  • HTML/Google Docs export (.html, .htm, .gdoc).
  • URL import (“reader mode” extraction via backend).
• Segmentation strategies:
  • Automatic detection (best strategy per document).
  • Markers (e.g. ---SEGMENT---, <!-- SEGMENT -->, ===).
  • Headings (H1/H2 hierarchy splitting).
  • Paragraph splitting and length-aware fallback splitting.

• Adaptive Comprehension Mode / Smart Import: AI analyzes text complexity and suggests:
  • better segment boundaries,
  • pacing adjustments,
  • optional pauses,
  • style suggestions,
  • “emphasis” candidates.
• Asynchronous job model with progress and result retrieval.

• Azure voice discovery (/api/voices) and caching.
• Voice preview library and preview sync/status tooling.
• Voice Library UI:
  • search + filters (locale, gender, styles),
  • favorites,
  • collections,
  • presets (voice + style + prosody),
  • default voices per context (quick_read, feeds, studio) synced across devices.

• Projects list with search and sorting.
• Rename, delete, multi-select operations.
• “Prepare audio” for project playback (merge segments into a single playback file when needed).
• Auto-save / draft tracking behaviors in the Studio lifecycle.

• Export types:
  • Audio (MP3/WAV).
  • Timeline JSON (project data).
  • Segment bundles (ZIP of per-segment assets, when applicable).
  • AAF bundle (ZIP containing an AAF-like XML + audio assets for import into editors).
• Export job history with progress tracking, downloads, and cleanup.

EchoLive Feeds is a full feed reader + ingestion system, not just an “RSS importer”.

• Subscribe/import:
  • RSS/Atom feed discovery from a website URL (handles direct-feed URLs).
  • Crawl-based “feeds” for non‑RSS sites (discover URLs and import selected pages).
  • Sitemap parsing for bulk URL discovery.
  • Curated recommendations and categories.
• Feed categorization:
  • Feed-level category (article | podcast | newsletter) with user override.
  • Item-level content type (article | podcast_episode | newsletter).
  • Podcast items play enclosure audio (no TTS); newsletters force TTS behavior.
• Feed settings:
  • folder organization,
  • fetch interval,
  • auto-generate audio (default voice),
  • cleanup patterns (preview/apply to remove repeated boilerplate).
• Items:
  • pagination, sorting, date filters,
  • search (title/excerpt/author),
  • read/unread, starred, bulk delete,
  • edit publication dates (single + bulk mappings).
• Reader experience:
  • full-screen Article Viewer (reading view) with font size controls,
  • refresh/re-fetch item content,
  • in-view audio controls + generation controls,
  • word-sync overlay for “read along” playback when timing is available.
• Audio for items:
  • per-item generate/regenerate/cancel with progress,
  • bulk generation with rate-limited queueing,
  • sticky player with seek + playback speed,
  • playback position persistence to backend.
• Scheduled refresh:

• Subscribe to YouTube channels via:
  • RSS conversion when available, and/or
  • YouTube Data API browsing (optional YOUTUBE_API_KEY) to go beyond typical RSS limits.
• Browse/search channel videos, import selected videos into a feed.
• Transcript + summary pipeline (optional):
  • fetch transcripts when available,
  • AI summarization,
  • optional summary-to-audio generation.
• YouTube playback mode in the global player (separate from TTS audio playback).

• In-app notification center backed by the database.
• Notification actions for common flows (e.g., retry from “insufficient credits” events).

• Credit-based prepaid billing model:
  • cost estimation by text length,
  • reserve credits on job start,
  • confirm on success,
  • refund on failure/cancel.
• Transactions + usage history.
• Coupons (validate/redeem) and admin tools for managing credits/coupons and cleanup/backfills.
• Credit expiry:
  • credit grants/purchases are issued with an expires_at timestamp (default 365 days),
  • the balance API can report credits expiring soon and the next expiration date,
• Subscription billing is planned; current model is pay-as-you-go via credits.

• Ingest → index → vector search pipeline over your own content:
• UI:
  • Cmd/Ctrl+K global search modal with keyboard navigation,
  • filter chips (All / Feeds / Projects).

• V2 parallel chunk-based synthesis with real-time progress streaming.

• Job progress polling + SSE streaming (where available).
• Defensive client UX: monotonic progress tracking, timeouts, resumable sessions.

• Encryption for sensitive user content at rest (project text, comprehension inputs/results).
• “No content logging” posture for user text/audio inputs; client monitoring redacts sensitive fields.
• GDPR endpoints for data export and account deletion.

• Pronunciation dictionary (per account/project), phoneme tooling, and reusable style/preset packs.
• Audio enhancement: background music mixing, fades/crossfades, loudness normalization/mastering.

• Pronunciation dictionary (per account/project), phoneme tooling, and reusable style/preset packs.
• Audio enhancement: background music mixing, fades/crossfades, loudness normalization/mastering.
• Collaboration: sharing, team workspaces, comments, version history.
• Integration/export expansion: richer DAW/video editor interchange, webhooks, public API.

• Voice cloning / custom voices (subject to provider constraints and policy).
• Multi-speaker scripting improvements and automated “role casting”.

• A mobile-first companion app focused on capture + offline listening (not a web wrapper):
  • camera/OCR → TTS,
  • share-sheet URL/article import,
  • background audio + downloads,
  • CarPlay/Android Auto,
  • lightweight editing + sync.

• More advanced summarization and “listenability” transformations (e.g., newsletter → concise audio brief).
• Semantic navigation and highlights over long audio (“jump to section”, “key points”).

• EchoLive is workflow-first: timeline/segment editing, bulk regeneration, exports, and long-form project management.
• EchoLive is ingestion-first: built-in feeds (RSS/crawl/newsletters/podcasts) and optional YouTube channel workflows.
• EchoLive is provider-leveraging: built around Azure Neural TTS (breadth of voices, regional availability) rather than proprietary voice models.
• Voice cloning/custom voice creation is not currently the core shipped differentiator (it’s a possible roadmap item).

• EchoLive is a studio + inbox (creation + production + export), not only a playback app.
• EchoLive has production controls (SSML, per-segment voices, AAF exports) that are typically out of scope for consumer readers.
• EchoLive’s Feeds model supports mixed media (articles + podcasts + YouTube) with unified playback and generation pipelines.
• EchoLive adds semantic search over your own library (ingest → index → vector search), which turns “saved content” into a searchable knowledge base.

• EchoLive generates high-quality audio from text with TTS-native controls and then can export to editing workflows (e.g., AAF bundle).
• EchoLive is not primarily a recorded-audio editor; it’s optimized for script → voiced audio pipelines.