# Coming Soon

The honest runway: what's still being finished in CoreReflex, what each item will do for you, and exactly what works today in the meantime.

**Shape this list:** [Feedback, Support & Roadmap](feedback-and-roadmap.md) covers chat bugs, feature requests, generation votes, and (shipping) public **upvotes**.

**On this page**

- [What changed](#what-changed)
- [Recently shipped · now in the rest of the guide](#recently-shipped--now-in-the-rest-of-the-guide)
- [How to read this page](#how-to-read-this-page)
- [The runway](#the-runway)
  - [CoreReflex's own generation models](#corereflexs-own-generation-models)
  - [Persona talking-head renders](#persona-talking-head-renders--the-lip-sync-engine)
  - [Richer CoreReflex Voice](#richer-corereflex-voice)
  - [Ingest public video & long-to-shorts 2.0](#ingest-public-video--long-to-shorts-20)
  - [Feature board & upvotes](#feature-board--upvotes)
  - [Email & SMS actions in Automations](#email--sms-actions-in-automations)
  - [Direct social account connections](#direct-social-account-connections)
  - [Hour-plus sources and in-page story controls in Auto-Clips](#hour-plus-sources-and-in-page-story-controls-in-auto-clips)
  - [A Files panel inside the editor](#a-files-panel-inside-the-editor)
  - [Your theme, saved to your workspace everywhere](#your-theme-saved-to-your-workspace-everywhere)
  - [Higher-level AI animation commands](#higher-level-ai-animation-commands)
  - [Desktop capture into Cut](#desktop-capture-into-cut)
- [Roadmap at a glance](#roadmap-at-a-glance)
- [A note on honest gating](#a-note-on-honest-gating)
- [Related pages](#related-pages)

---

## What changed

Earlier versions of this page were built around one big pending item: the self-hosted render worker. **That's done.** Rendering now runs end to end in production: queue a render from the Render Studio (or finish an Auto-Clips run, or export a short) and a finished, downloadable file lands in your library, baked with your grade, transitions, captions, and audio effects. There is no "provisioning" message left on that path.

Since then, a large batch of features shipped and came off this page entirely, around thirty of them, from auto-captions and transcript editing to the Auto-Clips studio, the Files section, in-editor comments, and the theme switcher. They're listed below with pointers to where each one now lives, so this page can go back to being what it should be: a short, honest list of what's *not* finished yet.

---

## Recently shipped · now in the rest of the guide

Everything in this section is live. Where a guide page covers it, the link takes you there; the rest is labeled with where to find it in the product.

**Render & delivery**

| Feature | Where it is now |
|---|---|
| Self-hosted rendering, end to end: queue a film in the **Render Studio**, get a finished MP4 (or WebM) master in your library | [Editing & Mastering](editing-and-mastering.md) |
| Shorts final export: a reframed 9:16 / 1:1 short renders on the same live pipeline | [Editing & Mastering](editing-and-mastering.md) |
| Platform export presets for YouTube, Shorts, TikTok, Reels, square, and X, with codec and aspect handled for you | Export panel in the editor |

**Auto-Clips: a whole new studio**

| Feature | Where it is now |
|---|---|
| Highlight detection on raw, long, unedited video: the AI watches the whole thing and returns ranked, non-overlapping moments with titles and hooks | **Auto-Clips** card on your dashboard, or [corereflex.com/auto-clips](https://corereflex.com/auto-clips) |
| Auto-cut vertical shorts: pick the moments you like and each one is trimmed, reframed to 9:16, rendered, and dropped into your library | Same page: prepare → detect → select → generate |
| Styled caption generation with nine caption personas (dramatic, funny, minimal, and more, plus auto-match to the moment) burned into the short | Same page, the captions step |
| Story mode + AI voiceover: one narrative arc across all your clips, narrated in a style-matched voice and auto-ducked under the source audio | Live via the platform API today; in-page controls are on the runway below |

**Captions & transcript editing**

| Feature | Where it is now |
|---|---|
| Auto-captions with word-level timing, style presets, and word animations (pop / rise) | [Editing & Mastering](editing-and-mastering.md) |
| Transcript panel: click a word to jump there, select words to delete them from the video, one-click remove filler words and long pauses | Transcript panel in the editor |
| Caption translation | Captions section in the editor |
| SRT + WebVTT subtitle export | Export panel in the editor |
| Find highlights: the AI ranks the best moments in a transcript; "Keep only this" cuts everything else in one undoable step | Transcript panel in the editor |

**Video tools**

| Feature | Where it is now |
|---|---|
| Green screen (chroma key) with similarity, softness, and spill controls | Inspector, "Green screen" section |
| Speed ramps (ramp in/out, slow-mo hit, rush presets) | Inspector, under the playback-rate slider |
| Extend with AI, continuing a clip from its own last frame | Inspector, Video section |
| Punch in: match ends (start+end frames), Transition (generate + hang a cut), Zoom (Ken Burns) | Inspector, Punch in section · ⌘K |
| Auto-reframe to canvas: the AI finds the subject and fits the crop | Inspector, Video section |

**Image & design**

| Feature | Where it is now |
|---|---|
| Generative fill, object erase, and expand (outpaint) with a mask brush | Image Studio canvas |
| Per-image adjustments and "Ask the AI" conversational edits | Image Studio canvas |
| Upscale 2× on any image node | Image Studio canvas |
| Multi-format export (PNG / JPG / WebP / SVG / PDF) and SVG import | Export menu in Image Studio and Slides |
| Logo studio, producing distinct concept batches with a 4-up treatment view (color / grayscale / stamp / reverse) and true vector export | Brand and logo flows on the dashboard |

**Slides**

| Feature | Where it is now |
|---|---|
| PowerPoint (.pptx) export | Export menu in Slides |
| Presenter mode with speaker notes and keyboard navigation | Present button in Slides |

**Audio**

| Feature | Where it is now |
|---|---|
| Mix room: multitrack sessions with mixer, clip editor, automation, record, bounce to program | Mix room in the editor; see also [Recording, Meetings & Podcasts](recording-and-podcasts.md) · [Voiceover & Music](voiceover-and-music.md) |
| Bounce to a mastered file with broadcast loudness targets, plus one-click auto-duck of music under voice | Mix room session toolbar |

**Timeline & editing**

| Feature | Where it is now |
|---|---|
| Keyframe animation with a per-segment curve editor (draggable bezier handles, presets) | Inspector, "Curve" tab; [Editing & Mastering](editing-and-mastering.md) |
| Advanced trims: Option-drag a clip edge for a ripple trim, Option-drag the body to slip, Option-Shift-drag to slide, and Q / W to trim to the playhead | Timeline gestures and shortcuts |
| Markers & chapters, including flags on the ruler, keyboard shortcuts, and a markers panel to list, jump, rename, and recolor | Composition inspector |
| Frame-accurate timecode grid and musical BPM grid, shared across video and audio | Grid controls in the editor |
| Version history: every save is snapshotted; restore any earlier version | Version-history panel in the editor |
| In-editor comments with @mentions, pinned to a timecode or an element | Comments panel in the editor |

**Review, library & provenance**

| Feature | Where it is now |
|---|---|
| Review-link upgrades: hover-scrub the video, compare versions side by side, draw directly on a frame to leave a note | [Proofing & Approvals](proofing-and-approvals.md) |
| Files, one home for everything you've uploaded and generated, with type filters, search that understands meaning ("sunset drone shot"), collections (folders), previews, download, and delete | **Files** in the dashboard nav |
| Content credentials: a signed provenance record on generated media, with a public verify check and an AI-generated flag | Badge on assets in the dashboard |

**Distribution & client operations**

| Feature | Where it is now |
|---|---|
| Social publishing: connect your GoHighLevel account once, then publish or schedule a render to Facebook, Instagram, LinkedIn, X, Google Business, and TikTok | **Publishing** view in the dashboard; [Client Operations](client-operations.md) |
| New automation triggers, **on new contact** and **on a schedule**, now fire alongside booking, meeting brief, payment, render done, and campaign approval | [Client Operations](client-operations.md) |
| New automation actions that publish a render to social, or run a whole creative workflow recipe on a schedule | Automation builder in the dashboard |
| Bulk create from CSV: one workflow recipe, one spreadsheet, one batch of finished assets | Bulk CSV flow in the dashboard |

**Studio-wide**

| Feature | Where it is now |
|---|---|
| Theme switcher: three presets plus five custom brand colors, applied across the site, dashboard, and editor with no white flash | Palette button on app pages |
| Diagram Studio, a branded whiteboard mode for sketches and flowcharts, exportable to your media library | Diagram mode in the editor |
| Workflow Studio: see your creative pipeline as a live node graph, or build a custom pipeline from typed stages with review gates | Workflow mode in the editor |
| AI-programmable editing: the assistant can drive real edits, including animation ("make the title fade in") | Assistant and JSON console in the editor |
| Chat support, an on-page assistant grounded in the docs, pricing, and guides | Chat bubble on the site |
| Human handoff + support tickets from chat | Same widget: **Talk to a human** / **Create ticket**. See [Feedback & roadmap](feedback-and-roadmap.md) |
| Generation thumbs up/down (quality routing + your workspace style memory) | Where shown on generated video/image results |
| Self-healing bug loop (ticket → structured repair brief → human-reviewed PR) | Backend for **bug** tickets; you never merge code. See [Feedback & roadmap](feedback-and-roadmap.md#how-feedback-makes-the-product-self-healing) |
| Cut (speech-native talk-to-timeline), HyperFrames films, web-grounded research films | Dashboard workflows + user guide pillars |

---

## How to read this page

Every runway item below has the same three parts:

- **What it is**: the plain-English description.
- **What it does for you**: why it's worth waiting for.
- **Status**: exactly what's live today versus what's shipping next.

For everything that already works, see the rest of the [user guide](README.md).

---

## The runway

### CoreReflex's own generation models

**What it is.** CoreReflex's own **image**, **video**, and **talking-avatar** generation models, running on CoreReflex's own generation hardware, including **long-form video** that can continue a scene for minutes, rather than only short clips you stitch together by hand.

**What it does for you.** The same kinds of generation you do today, at a **much lower credit cost**, on infrastructure CoreReflex owns end to end: the same "our own, far cheaper" trade you already get with **CoreReflex Voice** for narration (see [Voiceover & Music](voiceover-and-music.md)). Long-form continuation, in particular, opens up minutes-long generated video without hand-chaining a dozen short takes.

**Status.** Today, premium generation is **live and excellent**: video and images run on top-tier vendor models (Veo for video, Imagen for images), so nothing is waiting on this to make great work right now. CoreReflex's own lower-cost models are **rolling out as the dedicated generation hardware (GPUs) comes online.** As each one lights up for your deployment, it appears as a **choice** you can pick for cheaper credits, exactly the way the voice picker already lets you choose CoreReflex Voice over the premium Studio HD voices. Until a CoreReflex model is available on your instance, the premium vendor model simply handles the job; nothing about your workflow changes except, eventually, the price. The **talking-avatar** side of this is tracked separately just below, as the persona lip-sync engine.

### Persona talking-head renders · the lip-sync engine

**What it is.** The final step of the persona pipeline: a rendered video of your consented clone actually *speaking* your script, lips matched to the audio. Two styles, one flow: **lip-sync** a real reference (a photo or a short clip) to your new script, or bring a **single still to life** as a natural talking head with its own head movement and expression. You pick fidelity: a fast, real-time-quality render for volume, or a higher-fidelity render for hero moments where likeness has to be perfect.

**What it does for you.** Record yourself once, then put yourself on camera as often as you need without a lens. Update a product line, localize a message, or refresh a seasonal pitch by changing the words, not by re-shooting.

**Status.** Everything around the final render is live: the **Persona Studio** in your dashboard creates personas, attaches reference media, runs the hard consent grant/revoke flow, and clones a voice you have the rights to use. See [Voiceover & Music](voiceover-and-music.md). The rendering engines that bake the talking-head video are now **built and self-hosted on CoreReflex's own infrastructure** (both the lip-sync and the still-to-avatar styles above). The last step is the **dedicated generation hardware coming online** for your deployment, the same GPU rollout that lights up CoreReflex's own image and video models just above. Until it's switched on, requesting a persona clip returns a clear "not configured yet" message instead of a fake result, never a fabricated video. Personas are a **Pro** capability; see [Plans & Billing](plans-and-billing.md).

### Richer CoreReflex Voice

**What it is.** More ways to narrate and clone on CoreReflex's own voice stack: broader languages, expressive delivery (emphasis and emotion on the line), and faster clone paths from a short sample, still chosen as **CoreReflex Voice** options in the picker, not a pile of third-party brand names.

**What it does for you.** Faceless and presenter films sound less generic; localization and brand reads get easier without jumping to a separate TTS vendor for every language.

**Status.** CoreReflex Voice (fast narration) and consented cloning paths already ship for many workflows; see [Voiceover & Music](voiceover-and-music.md). The next wave expands language coverage, expressive controls, and clone quality tiers as each engine is brought online and quality-checked. Until a tier is available on your deployment, the picker simply shows what works today.

### Ingest public video & long-to-shorts 2.0

**What it is.** Paste an allowed public video URL into CoreReflex (where the platform's policy permits), land it in your library, then run **Cut** or a stronger **long → ranked vertical shorts** factory, plus optional “make something like this, but about *my* topic” briefs that keep pacing and structure without copying the original.

**What it does for you.** Less downloading and re-uploading; faster repurposing of webinars, podcasts, and public references into editable projects and social cuts.

**Status.** Auto-Clips and Cut already handle uploads and many URL prepares; full **policy-gated public ingest**, reference-style briefs, and the next clip-factory ranking pass are on the runway. Nothing will pull private or disallowed sources.

### Feature board & upvotes

**What it is.** A signed-in **feature board**: browse requests, **upvote** what you need most, submit a short request, and see status move from *under consideration* → *planned* → *in progress* → *shipped* (with a link to the guide when it lands).

**What it does for you.** Transparent priority. One vote per account per item. Your chat feature requests today already go to the team; the board makes demand visible to everyone.

**Status.** **Live:** public board at **[/roadmap](/roadmap)** (list, submit when signed in, one upvote per account), chat **Report a bug** / **Request a feature**, generation thumbs, and this narrative runway. Staff can change status via God Mode admin APIs. Apply migration `054` on the API database if the board is empty. Upvotes inform priority; they do not force unsafe or illegal work onto the roadmap.

### Email & SMS actions in Automations

**What it is.** New action types for the automation builder that send an email or a text message directly, plus **wait steps** so an automation can pause ("wait two days, then follow up") instead of running everything at once.

**What it does for you.** Today, "a booking came in" can create a task, log activity, advance a campaign, call a webhook, run a workflow recipe, or publish to social. But if you want it to *message someone*, you have to route through a webhook to an outside tool. Native email/SMS actions close that loop inside CoreReflex, on your own brand.

**Status.** The automation engine, builder, and seven triggers are live (see [Client Operations](client-operations.md)); every automation currently runs its actions immediately, in order. Sending email/SMS from an automation, delay/wait steps, and multi-step drip sequences are the next layer of the engine.

### Direct social account connections

**What it is.** Native, per-platform sign-in for social publishing: connect Facebook, Instagram, LinkedIn, X, Google Business, or TikTok to CoreReflex directly.

**What it does for you.** One fewer account in the chain, and publishing that doesn't depend on any relay service.

**Status.** Social publishing is **live today via a connected GoHighLevel account**: connect it once in the dashboard's **Publishing** view (your key is stored encrypted), then publish or schedule any render to your connected platforms, with every attempt logged so a failed post is never silent. Direct per-platform connections are the v2 upgrade behind the same publishing screen. When they land, nothing about your workflow changes except the connect step.

### Hour-plus sources and in-page story controls in Auto-Clips

**What it is.** Two upgrades to the Auto-Clips studio: support for very long (hour-plus) source videos, and on-page controls for story mode, AI voiceover, and picking a source straight from your media library.

**What it does for you.** Drop a full-length podcast, stream VOD, or webinar recording in and get ranked vertical shorts out, and drive the narrated "story" version of your clips without touching the API.

**Status.** Auto-Clips is live end to end: paste a video URL, **Prepare** builds a compact analysis copy so long videos can be understood in full, **Detect** returns ranked moments, and **Generate** renders your selected moments as captioned vertical shorts into the library. Two edges remain. First, on *very* long sources (roughly hour-plus), even the compact analysis copy can exceed the analysis size limit. A staging upgrade for those is planned; if you hit it, split the source or start from a shorter cut. Second, **story mode and style-matched voiceover work today through the platform API** but aren't yet buttons on the Auto-Clips page, and the page takes a pasted URL rather than a picker over your media library. Both are UI work on a shipped backend.

### A Files panel inside the editor

**What it is.** A Files tab inside the studio editor that browses your whole workspace library (everything you see in the dashboard's **Files** section) and imports any of it into the open project with one click.

**What it does for you.** No more round-tripping to the dashboard to find last month's upload or a render from another project; your entire library becomes source material for whatever you're editing.

**Status.** The **Files** section itself is live in the dashboard: every upload and generated asset in one place, with type filters, meaning-aware search, collections, previews, download, and delete. The editor currently shows only the assets already in your open project. The workspace-to-project bridge panel is the next piece.

### Your theme, saved to your workspace everywhere

**What it is.** The white-label theme you pick (preset plus custom brand colors), saved as your **workspace's** default, so every member sees your brand look on any device, plus a full Appearance panel in Settings and a theme switcher inside the editor.

**What it does for you.** Brand the whole suite once (site, dashboard, and editor) and have it follow your team everywhere, instead of living in each person's browser.

**Status.** The switcher is live on app pages (the palette button): three presets, five custom brand colors, applied instantly with no white flash, remembered per device, and carried across into the editor. Saving the theme as the workspace-wide default is built and rolling out. Until it's enabled on your deployment, the workspace save may report that theme persistence isn't available yet, and your choice stays device-local. Still ahead: a proper **Appearance** panel in dashboard Settings (more tokens, logo, fonts) and a switcher you can open from inside the editor.

### Higher-level AI animation commands

**What it is.** Motion described in one sentence ("typewriter, then bounce on the beat"), turned into a complete, editable keyframe animation: whole-clip animation styles, generated keyframe sets, and staggered multi-item motion.

**What it does for you.** Animation quality without animation labor. You describe the feel; the assistant programs the keyframes; everything it writes stays a normal, hand-editable animation on the curve editor.

**Status.** The foundation is live: the assistant can already program animation directly, setting keyframes, applying easing (including custom bezier curves), enabling or clearing animation, and moving items. That means "make the title fade in" works today, and everything it authors renders identically in preview and export. The next layer is the higher-level vocabulary: generate a full keyframe set from a description, animate a clip in a named style, and stagger an animation across several items.

### Desktop capture into Cut

**What it is.** Capture system or meeting audio (and related desktop flows) and send it straight into CoreReflex **Cut** (speech-native edit, captions, editable project) without a manual file shuffle.

**What it does for you.** Meeting → cleaned cut, podcast → tight edit, less friction between capture and the talk-to-timeline wedge.

**Status.** Meeting import and related capture bridges are rolling out; a full desktop shell for system audio is next on the capture runway. Until your deployment shows the capture path, use upload + Cut as today.

---

## Roadmap at a glance

| Capability | Live today | Shipping next |
|---|---|---|
| **Generation models** | Premium vendor models (Veo for video, Imagen for images) live; CoreReflex Voice live for narration | Lower-cost CoreReflex Image, Video, long-form video, and talking-avatar as generation GPUs come online |
| **Persona clips** | Persona Studio: create, reference media, consent grant/revoke, voice cloning | Self-hosted lip-sync / talking-head final render on generation hardware |
| **CoreReflex Voice** | Own narration + clone paths in the picker for many workflows | More languages, expressive delivery, higher-quality clone tiers |
| **Ingest & repurpose** | Upload + Auto-Clips + Cut | Policy-gated public URL ingest, reference-style briefs, clip factory 2.0 |
| **You → product** | Chat Q&A, human handoff, tickets, bug→repair brief, gen thumbs | Public **feature board** + upvotes + status badges |
| **Automations** | 7 triggers, 6 actions (tasks, activity, campaigns, webhooks, recipes, social) | Email/SMS send actions, wait/delay steps, drip sequences |
| **Social publishing** | Publish/schedule to 6 platforms via a connected GoHighLevel account | Native per-platform account connections (no relay) |
| **Auto-Clips** | Long video → ranked moments → captioned vertical shorts in the library | Hour-plus source staging; story + voiceover controls and a library picker on the page |
| **Files** | Dashboard Files section: filters, semantic search, collections | A Files panel inside the editor with one-click import into the open project |
| **Theming** | Presets + custom colors, per device, across site/dashboard/editor | Workspace-wide default everywhere, Appearance panel in Settings, in-editor switcher |
| **AI animation** | Assistant-programmed keyframes, easing, fades ("make the title fade in") | One-sentence motion: generated keyframe sets, styles, staggered animation |
| **Capture** | Meeting/import paths rolling out | Desktop system-audio shell → Cut |

Everything else in the product (directing, producing, voiceover, scoring, recording, captions, transcript editing, image AI editing, slides, audio sessions, rendering, shorts, Auto-Clips, files, comments, versions, proofing, publishing, client ops) runs end to end today. See [Recently shipped](#recently-shipped--now-in-the-rest-of-the-guide) above and the [user guide](README.md).

---

## A note on honest gating

Some capabilities on this page are "engine-gated": the studio surface is real and open, but a heavy-lifting engine behind it isn't switched on for your deployment yet. In every one of those cases, CoreReflex tells you so plainly: a persona clip without the lip-sync engine, or a workspace theme save before persistence is enabled, returns a clear, actionable message instead of a silent hang or a faked result. If a button exists, it either works or it says exactly why it can't yet.

Have a capability you're especially waiting on? **[Tell us in chat](feedback-and-roadmap.md)**. Bugs open a repair path; feature requests and (soon) **upvotes** shape what ships first.

---

## Related pages

- [Feedback, Support & Roadmap](feedback-and-roadmap.md) for chat, bugs, votes, and the feature board.
- [User guide home](README.md), the full map of what's live.
- [Editing & Mastering](editing-and-mastering.md) for the timeline, captions, shorts, and rendering.
- [Voiceover & Music](voiceover-and-music.md) for narration, scoring, and consented voice cloning.
- [Client Operations](client-operations.md): CRM, booking, tasks, automations, and publishing.
- [Proofing & Approvals](proofing-and-approvals.md): client review links.
- [Plans & Billing](plans-and-billing.md): what each plan unlocks.