Roadmap

See what we’re building next.

We’re building VelaClip in layers. Reliable editing foundations first, church-specific intelligence next, and automation only after the workflow earns trust.

View what’s in progressSee what shipped
Sourced from LinearPublic prioritiesSubject to change
Sequence
Foundations → autonomy, in order.
  1. FoundationsActive
  2. Core Intelligence
  3. Differentiated Workflows
  4. Scale & Automation
  5. Platform Expansion
Philosophy

We’re not building every AI feature at once.

VelaClip’s roadmap is deliberately sequenced. Editing primitives come before agent autonomy. Sermon understanding comes before archive-wide automation. Review and reversibility come before recurring background actions.

  1. Principle 01

    Reliable before autonomous.

    AI should not silently mutate a project before every action can be previewed, explained, and reversed.

  2. Principle 02

    Church-specific before generic.

    Scripture, testimony, prayer, sermon structure, and message integrity deserve dedicated intelligence — not a generic clip picker.

  3. Principle 03

    Automation after trust.

    Recurring workflows come after the single-project experience is stable, explainable, and worth trusting.

Active milestone

Foundations

Stabilize the architecture and core editing primitives that unlock the intelligence and automation layers.

Now · Foundations
Weeks 1–2
Target · Jul 19, 2026
Progress0 of 11 complete · 0%
Exit criteria

Typed actions, structured plans, resumable workflows, scene-aware analysis, and faster local rendering.

Committed work
  • Typed Tool Registry
    Included
    In progress
  • Edit Plan object + review UI
    Included
    In progress
  • Scene detection
    Included
    In progress
  • Transcript editing modes
    Free
    Committed
  • Separate media/audio tracks
    Free
    Committed
  • Transcription backend benchmark
    Included
    In progress
  • Granular progress and cancel UX
    Free
    Committed
  • Editor navigation consolidation
    Free
    Committed
  • Home workspace redesign
    Free
    Committed
  • Multi-service Make Clips queue
    Free
    Committed
  • VideoToolbox acceleration
    Free
    Committed
Overview

Now, next, and later — at a glance.

A shorter view of the same roadmap. Deeper phase details continue below.

Now

Actively being built or committed in the current execution window.

Next

Planned after current foundations and intelligence work.

Later

Valuable work intentionally deferred until dependencies land or validation is stronger.

Detailed roadmap

Every phase, curated.

Public feature groups instead of raw issue dumps. Expand any phase for the underlying work, with its live status pulled from planning.

Now · Foundations

Foundations

Weeks 1–2Target · Jul 19, 20260/11 shipped
Copy link to phase

Stabilize the architecture and core editing primitives that unlock the intelligence and automation layers.

Dependency

Everything downstream — Vela Orchestrator, Plan → Preview → Apply, Vela Edit — depends on these primitives landing first.

Safer AI editing foundations

A typed internal tool layer and structured Edit Plans so every AI action can be inspected, previewed, and reversed.

  • Typed Tool Registry
  • Edit Plan object + review UI
Stronger editor primitives

Cleaner ways to work with transcripts, tracks, and scene boundaries.

  • Transcript editing modes
  • Separate media/audio tracks
  • Scene detection
Faster local transcription on Mac

Benchmarking transcription backends for speed and accuracy on Apple Silicon.

  • Transcription backend benchmark
Better Mac experience

Granular progress, a redesigned home workspace, consolidated navigation, VideoToolbox acceleration, and multi-service queues.

  • Granular progress and cancel UX
  • Editor navigation consolidation
  • Home workspace redesign
  • Multi-service Make Clips queue
  • VideoToolbox acceleration
Now · Core Intelligence

Core Intelligence

Weeks 3–6Target · Aug 16, 20260/11 shipped
Copy link to phase

Build differentiated intelligence on top of the stabilized primitives — sermon understanding, scripture, reversible agent edits, and better audio.

Dependency

Requires Foundations. Sermon Intelligence and Scene Detection unblock narrative clip selection, context guard, and smart reframe.

Sermon Intelligence

A local model of the sermon — structure, scripture, testimony, and the message being taught.

  • Sermon Intelligence Graph
  • Scripture detection and verification
Narrative clip selection

Rank moments by how they land in the message, not just by keyword density.

  • Narrative-aware candidate generation
  • Local intent-aware scoring
Context Guard

Detect and warn when a clip risks stripping a sermon moment of its context.

  • Message Integrity + Context Guard
Reversible agent editing

An orchestrator that proposes edits as Plans you preview and apply, never silent mutations.

  • Vela Orchestrator
  • Plan → Preview → Apply change sets
Visual and audio intelligence

Active-speaker reframing, Enhance Voice, and church-aware proofreading for captions.

  • Active-speaker smart reframe
  • Enhance Voice pipeline
  • Social pacing, filler, and pause tools
  • Church-aware proofreader
Now · Differentiated Workflows

Differentiated Workflows

Weeks 7–10Target · Sep 13, 20260/10 shipped
Copy link to phase

Ship end-to-end workflows that only make sense once the primitives and church intelligence are in place.

Dependency

Vela Edit V1 integrates Edit Plans, clip intelligence, Context Guard, reversible changes, scene detection, smart reframe, Enhance Voice, and social pacing.

Vela Edit V1

Reviewable, reversible AI-assisted editing across a full sermon.

  • Vela Edit V1
Sermon-to-Content Week

One sermon planned as a coordinated week of shorts, captions, and posts.

  • Sermon-to-Content Week
Archive intelligence

Church Media Memory and semantic search across your sermon archive.

  • Church Media Memory
  • Cross-archive semantic search
Smart B-roll Director

Contextual B-roll suggestions drawn from your own media library.

  • Smart B-roll Director
Hook and variant studio

Hook Laboratory, branching variations, packaging, and duration variants.

  • Hook Laboratory
  • Branching and variations
  • Automatic clip packaging
  • Duration and narrative variants
Performance intelligence

Explainable insight into which clips resonated and why.

  • Explainable Performance Intelligence
Next · Scale & Automation

Scale & Automation

Target · Nov 15, 20260/15 shipped
Copy link to phase

Once single-project editing is trustworthy, extend into automation, multi-clip assembly, multilingual reach, and channel autopilot.

Dependency

Blocks on Differentiated Workflows landing. Nothing here should ship before the review-and-approve editing loop earns trust.

Beyond a single clip

Multi-clip AI assembly, per-shot effects, and specialist agents that work together.

  • Multi-clip AI assembly and merge
  • Per-shot AI effects
  • Specialist agents
Audio and multilingual

Stem separation, translated captions, and dubbing for multilingual ministries.

  • Audio stem separation
  • Multilingual captions, translation, dubbing
Church-specific styling

Edit styles, reframing modes, and safe-zone intelligence tuned to church content.

  • Church-specific AI Edit Styles
  • Church-aware reframing modes
  • Auto-zoom and safe-zone intelligence
Content operations

Performance-learned selection, AI scheduling, Watch Folders, and YouTube Channel Autopilot.

  • Performance-learned selection
  • AI content strategy scheduling
  • Watch Folder automation
  • YouTube Channel Autopilot
Durability and onboarding

Resumable workflows, intent-driven onboarding, and a safer external automation API.

  • Durable resumable workflows
  • Intent and role onboarding
  • Safe external automation API
Later · Platform Expansion

Platform Expansion

0/5 shipped
Copy link to phase

Directions worth pursuing once the automation layer is proven — deep archive agents, learned church style, and broader interoperability.

Dependency

We will only pull these forward when they directly unblock a validated user need or materially improve retention.

Deep archive intelligence

Cross-sermon agents that reason across a full teaching archive.

  • Cross-sermon archive agent intelligence
Church-learned style

Editing and packaging tuned to your own history of shipped content.

  • Learned church style from history
Interop and remixing

Broader MCP exposure and advanced church-safe remixing.

  • Broad MCP exposureAdd-on
  • Advanced church-safe remixing
Optional generated media

Opt-in AI-generated visuals or motion, offered as an add-on.

  • Optional AI-generated media assetsAdd-on
Critical path

Some features only make sense after the foundations are stable.

The order of the roadmap is not arbitrary. These chains show which pieces have to land first before the more visible features become safe to ship.

Reversible AI edits

The typed tool layer and structured Edit Plans have to land before any agent can propose reviewable changes.

  1. Typed Tool Registry
  2. Edit Plans
  3. Plan → Preview → Apply
  4. Vela Orchestrator
Narrative clip selection

Ranking clips by how they land in the sermon requires both meaning and scene structure.

  1. Sermon Intelligence Graph
  2. Scene detection
  3. Narrative-aware selection
  4. Context Guard
Smart B-roll Director

Only useful once VelaClip understands the sermon and remembers your church's media.

  1. Sermon Intelligence
  2. Church Media Memory
  3. Smart B-roll Director
Vela Edit V1

The end-to-end AI edit integrates every foundation and intelligence primitive.

  1. Edit Plans
  2. Clip intelligence
  3. Context Guard
  4. Reversible changes
  5. Smart reframe + Enhance Voice
  6. Vela Edit V1
Recently shipped

Looking for what already shipped?

Real releases live in What’s New — verified version notes, dates, and detailed changes.

View What’s New
Status legend

Six honest statuses.

Every card on this page carries one of these. Cancelled and duplicate work is hidden by default.

  • In progress

    Actively being built right now.

  • Committed

    Committed to the active execution window.

  • Next

    Planned after the current milestones.

  • Later

    Intentionally deferred until validated or unblocked.

  • Blocked

    Waiting on a dependency or decision.

  • Shipped

    Already released — see What's New.

Feature request

A workflow missing from the roadmap?

Tell us what your church is trying to do. We care more about the real workflow than a generic feature request.

Share a feature request
Transparency

A roadmap is a direction, not a promise carved in stone.

Priorities can change as we learn from real churches, technical constraints, and product usage. We publish the roadmap to show what we’re focused on — not to pretend every date is guaranteed.

  • Dates are targets, not guarantees.
  • Dependencies may shift the order of committed work.
  • Shipped work belongs in What’s New, not on this page.
  • Unstarted work may change as we learn from real churches.
  • Customer feedback can raise or lower priority.
What’s next

Start with what VelaClip can do today. Follow where it’s going next.

View What’s NewExplore use cases
Local-first·Built for churches·Public roadmap