Features

All three creation modes support every format and duration. Set it once — the entire pipeline adapts: canvas size, caption positioning, thumbnail dimensions, and provider aspect ratios.

3 Video Formats

5 Duration Presets + Custom

Used in Single Video and knowledge-based Series modes. Episio fetches trending topics from multiple live sources and lets you choose or bring your own. Seven content domains focus the research:

Research sources: Web search, Reddit, RSS feeds, Hacker News — aggregated and ranked by recency and relevance. AI synthesizes findings into your script automatically.

Synthesis topics: AI-generated cross-domain angles that connect ideas from different fields — often the most viral format.

Synthesis topics ★ — A special topic type that connects ideas across domains. When selected, the Synthesis script style is automatically pre-set. These are often the highest-performing videos for virality because they surface non-obvious angles nobody else has covered.

Narrative angle analysis: After entering a topic, click "Analyze narrative angles" — AI generates 4–5 distinct narrative frames for the same topic. Each angle changes what the story is "about" while using the same facts. Pick the one that fits your channel voice.

Document upload: Upload a PDF, DOCX, TXT, or Markdown file instead of typing a topic. Episio extracts the text and uses it as the research source — great for turning reports, articles, your own notes, or transcripts into videos.

Scripts are generated by Claude AI. Two dimensions shape the output: script style (structure) and narrator type (voice and tone). 100 unique combinations.

10 Script Styles

10 Narrator Types

Script Review — Gate Before Footage

After generation, you land on the Script Review page before any footage is generated. This is where you:

• Read the full script and edit any sentence directly
• Refine via AI chat — "make the opening more dramatic", "add a statistic", "cut to 60 seconds"
• Switch script style or narrator type and regenerate
• Approve and proceed to footage generation only when satisfied

Episio doesn't pick music from a library. It reads the emotional arc of your script and composes an original soundtrack from scratch — matched exactly to your video's length. Original every time. No royalty issues. No ContentID strikes on YouTube.

8 emotional profiles — you set the direction; AI composes within it:

Auto-classification: AI reads your script and selects the emotional profile automatically. You can override it at any time.

Custom MP3 upload: Upload your own track — it replaces the composed score entirely. Good for branded channels with a signature sound.

Music remix: Change the profile after video generation and reassemble with one click.

Episio supports 17 voice languages for voiceover generation. Captions render natively in 10 languages (English + 5 Indian + 4 European). Select the language on the creation page — the voice engine auto-routes based on your choice.

ElevenLabs Voices

High-quality voices with distinct accents and tones.

Sarvam AI — Indian Languages

Native-quality voiceover for 5 Indian languages. Episio auto-routes to Sarvam AI when an Indian language is selected — no manual setup needed.

Voice Cloning

Upload a 30-second audio sample → Episio creates a clone via ElevenLabs. Use it across all your videos for a consistent personal brand voice. Set it as your default in Settings.

Custom Voice ID: Have an ElevenLabs voice ID not listed? Paste it directly — Episio supports any ElevenLabs voice ID.

14 AI providers power footage generation — all via your FAL key. The footage director plans each scene automatically, or you can override per scene.

Quality Tiers

Free sources: Pexels stock footage, Archive.org historical footage, NASA imagery — all at zero credit cost. The footage director balances free and AI sources automatically.

See the Provider Guide for full comparison with costs, render times, and recommended combos.

Scene Studio gives you full control over any scene in your video. Open it from the review page by clicking any scene card. 8 ways to create or replace a scene:

Footage Sources

Procedural Visuals — Zero Credit Cost

Scene Sub-clip Splitting

Any scene can be split into 2, 3, or 4 sub-clips. Each slot is filled independently — use AI generation for one, upload for another, or pull from Pexels. The narration text is divided proportionally across the slots. Ideal for long scenes that need visual variety.

Image upload → Ken Burns clip: Upload any JPG, PNG, or WebP image from Scene Studio and Episio applies Ken Burns motion (zoom in, zoom out, pan left, pan right at slow/normal/fast speed) to create a 6-second animated clip at zero AI credit cost.

Creative Series is a distinct creation mode for purely imaginative content. No topic research, no news — just a creative vision, AI-built characters, and episode after episode of cinematic storytelling.

What Makes It Different

The Brainstorm Drawer

A side drawer that opens while you write your episode scenario. The AI knows your full world bible and character list — every suggestion is grounded in your specific world.

Topics you can explore in the drawer:

• Camera anglesLow-angle handheld, wide establishing shot, tight close-up on a prop — specific and cinematic.
• World buildingRecurring visual motifs, sound design palette, era-specific details that anchor the setting.
• Character depthMotivations, relationships, unspoken conflicts, how a character would react to this scene.
• Scene ideasEpisode structure, what happens, bittersweet turns, how to show emotion without dialogue.
• DialogueLines characters speak in lipsync mode, or narration phrasing for ambient episodes.
• Mood & toneEmotional arc, pacing notes, how the score should shift through the episode.

Two Modes

What Kind of World Can You Build?

Scene Plan Review

After entering your episode scenario, AI generates a scene-by-scene plan before generating any footage. You see: scene number, duration, scene type, characters appearing, and a brief scene description. Review the plan, then click Generate Episode to commit. If the plan isn't right, regenerate with different scenario notes.

Knowledge-based Series (for topic-driven channels) use a world bible to maintain consistent visual identity and character continuity across episodes. Every series has a world bible — the creative DNA that controls how AI generates each episode.

World Bible Sections

Characters

Each character has: name, role (Protagonist / Deuteragonist / Antagonist / Ally / Supporting), visual description, AI-generated portrait, and a Veo3 hint — a short phrase (e.g. "tall man with gray beard") appended to every scene prompt that features this character. Approve one portrait as "canonical" — that face is locked for all episodes.

Shot Arc Phases

Group episodes into story arcs — each phase has its own camera language and optional character age. This lets you evolve the visual style as the story progresses:

• Shot style — camera language for this phase (e.g. "intimate close-ups, handheld" for early episodes; "wide epic shots, crane" for the climax)
• Default character age — set once per phase for biography series; AI references the age in scene prompts automatically
• Quality tier — Economy for setup, Standard for core, Premium for finale

Continuity Features

• AI suggests episode topics based on series context and prior episodes
• Continuity checker flags contradictions with previous episodes
• Character pre-flight check warns about missing portraits before generation
• World bible injected into every scene regeneration prompt

Create talking-head videos from a portrait photo or video clip. The avatar speaks your voiceover with lip-synced animation. Available on the Studio plan.

Two Input Modes

3 Lipsync Engines

Burned-in captions — no CapCut, no Descript, no subtitle files. Auto-generated from your voiceover with word-level timestamp sync. Configure on the review page before publishing.

Captions are burned into the final MP4 via FFmpeg — upload directly to YouTube, TikTok, or Instagram with no separate subtitle file needed.

📸

Caption configuration panel

Screenshot coming soon

AI-generated thumbnails via Ideogram — no Canva, no Photoshop, no designer. The prompt is auto-generated from your script and topic. Available in 3 styles:

4 AI variants at once: Click "Generate thumbnails" to produce 4 different variants simultaneously. Claude Vision scores each one for click-through potential — titles, contrast, emotional pull. The top-scored variant is pre-selected, but you can pick any.

Edit the prompt: Modify the generation prompt to steer the style and composition before generating.

Series thumbnails: Character portraits are injected into the thumbnail prompt for face-consistent thumbnails across all episodes.

The review page is where you fine-tune your video before publishing. You control every scene, the music, captions, and thumbnail — nothing is locked.

Scene Grid

Each scene card shows: clip preview, narration text, type badge (AI Video, Stock, Image, Chart, etc.), source provider, mood, and status. Click any scene to open Scene Studio.

Regenerating a Scene

Edit the prompt, select a different AI model, upload your own clip, or browse Pexels stock footage. Cost and ETA shown before confirming.

Reassembly

After editing scenes, music, or captions — click Reassemble to re-render the full video. Takes ~1–2 minutes.

📸

Review page with scene grid

Screenshot coming soon

Publish directly to YouTube from the review page. Connect your channel in Settings → Publishing.

YouTube Publish Panel

• Channel selector (if multiple channels connected)
• Title (100 character limit)
• Description
• Hashtags (auto-suggested from topic)
• Pinned comment
• Privacy: Private, Unlisted, or Public

Real-time status: uploading → processing → published. You get the YouTube link as soon as it's live. Can't publish now? Download the MP4 and post manually to TikTok or Instagram Reels.

All your videos in one place. Filter by status: All, Completed, Review Needed, In Progress, Failed.

Run cards show: thumbnail, topic, date, cost, and status badge. Active runs display a live progress bar with stage counter.

Actions: View/edit, download video + thumbnail, or publish — all from the library card.

API Keys

Connect your keys for each service. Each key has a test button to verify connectivity.

Brand Kit

• Logo watermark: Upload your logo (PNG/JPG ≤2MB). Choose corner position (4 options) and opacity (30–100%) — burned into every video automatically.
• Branded intro clip: Upload an MP4/MOV intro (≤200MB) — prepended to every video before the generated content. Great for channel identity.
• Default voice + language: Set once — applied to every new video. Change per-video at creation time.

Presenter / Avatar (HeyGen)

Settings → Presenter. Configure an avatar that appears in talking-head scenes as an alternative to lipsync:

• HeyGen avatar picker: Browse and search from hundreds of ready-made avatars. Choose rendering style (full body, close-up, circle).
• Photo avatar: Upload a portrait photo of yourself — Episio creates a personalized avatar via HeyGen (5–20 minute processing time). Your face, your avatar.
• HeyGen API key: Required for avatar mode — add in Settings → API Keys.

YouTube Connection

Settings → Publishing → Connect YouTube. OAuth flow links your channel for direct publishing. Multiple channels supported — choose per video.