solanasis-ai-assisted-youtube-video-production-playbook-research-handoff-2026-03-18

AI-Assisted YouTube Video Workflow Research Playbook, Briefing Memo, and Handoff for Another AI — 2026-03-18

Executive Summary

This document extracts, verifies, and improves the key conclusions from the prior discussion about building a mostly-AI-assisted YouTube workflow where Dmitri still does some of the talking on camera, keeps editing pain low, and stays roughly under $100/month in software.

The central conclusion still holds, and it is now better supported: the strongest pattern is not “one magic AI app does everything.” The most credible and practical creator workflow is usually:

1. Real talking-head footage as the backbone
2. Transcript-based editing for rough cuts
3. AI-generated mini-clips / B-roll only where they add value
4. Fast audio cleanup
5. Optional long-to-shorts automation

That conclusion is supported by both official product pages and current user-reported Reddit workflows, with an important caveat: user reports are anecdotal, uneven, and sometimes contradictory.

The best-value recommendation for this use case remains:

• Descript Creator for transcript-first editing
• Runway Standard or Pika Basic/Standard for short AI-generated visual inserts
• Adobe Podcast Premium or Auphonic for audio cleanup
• Optional: OpusClip Starter if the main goal includes turning long videos into Shorts

The most practical starting stack under budget is:

• Balanced stack: Descript Creator ($24/moannual),RunwayStandard($12/mo annual), Adobe Podcast Premium ($9.99/moor$99.99/year), for roughly $46/mo before tax
• Lean stack: Descript Hobbyist ($16/moannual),PikaBasic($8/mo annual), Auphonic Free (2 hours/month), for roughly $24/mo before tax

A major improvement over the prior discussion is that this document now explicitly adds the missing operational and policy constraints:

• YouTube requires disclosure for meaningfully altered or synthetic content when it seems realistic.
• YouTube monetization depends on originality, and reused/mass-produced-looking channels are at higher risk.
• AI clip generators are best used for short inserts, not as the primary engine of the full video.
• Audio quality is disproportionately important; better sound often matters more than more AI visuals.
• Human review is still required for pacing, narrative judgment, clip selection, and brand quality.

Purpose of This Document

This is intended to function as all of the following:

• a research-grade extraction of the original discussion,
• a verified decision memo,
• an implementation playbook,
• a briefing artifact for Dmitri / Solanasis,
• and a handoff document that another AI can use without reopening the original conversation.

It is deliberately more structured and more cautious than the original answer. It separates verified facts from recommendations, labels uncertain points, preserves links, and adds missing considerations that should have been included the first time.

Discussion Context

User goal

• User-stated: Dmitri wants a workflow for making YouTube videos that uses mostly AI tools, keeps him involved as the person speaking, produces a professional result, and avoids painful editing.
• User-stated: Budget target is roughly under $100/month, ideally well below that when possible.
• User-stated: The desired result is not a faceless fully-automated AI channel; the user wants real talking + AI support, especially through AI-generated mini videos / B-roll, stitching, captions, and cleanup.

Core question from the original discussion

• User-stated: Can Dmitri use relatively affordable AI tools to create polished YouTube videos with his own talking-head footage plus AI-generated mini clips, without spending huge time in editing?

Evidence Legend

Every important point in this document is tagged with one of the following statuses:

• Verified = supported by a current official source or by a directly checkable source; for Reddit items, this means the report exists, not that the claim is universally true.
• User-stated = came from the user’s request or preferences.
• Assistant-stated but unverified = recommendation or inference that was given previously but is not directly proven by a strong source.
• Tentative / speculative = plausible but uncertain, conflicting, or highly dependent on changing product behavior, credits, prompts, region, or creator skill.

Key Facts and Verified Findings

1) The strongest low-friction workflow pattern is transcript-first editing plus selective AI visuals

• Verified: Descript’s official pricing/features page confirms transcript-first, AI-assisted video editing features, including Underlord, Remove Filler Words, Studio Sound, Create Clips, and 4K export on Creator.
  Source: https://www.descript.com/pricing
• Verified: Riverside officially offers recording, text-based editing, AI repurposing, local recording, separate tracks, up to 4K video, and up to 48 kHz audio.
  Sources: https://riverside.com/ and https://marketing.riverside.fm/pricing
• Verified (user-reported anecdote): Recent Reddit discussions repeatedly describe transcript-first rough-cutting as the easiest way to reduce editing pain for talking-head videos.
  Source: https://www.reddit.com/r/youtubers/comments/1rqth58/i_finally_found_a_less_painful_way_to_rough_cut/

Interpretation:

• Assistant-stated but unverified: For Dmitri’s use case, transcript-first editing is likely the highest-leverage time saver because the painful part is usually not fancy effects, but getting from rambling raw takes to a usable first cut.

2) There is no credible “one-tool does everything perfectly” consensus

• Verified (user-reported anecdote): In a recent podcasting thread, one creator explicitly described a split workflow: “I record in Riverside, edit in Descript, clip in Opus,” and said Descript and Opus were better than doing it all inside Riverside.
  Source: https://www.reddit.com/r/podcasting/comments/1redo6x/is_riverside_business_worth_it_for_webinars/
• Verified (user-reported anecdote): Another current thread says AI text-based editing is strong for mechanical cleanup, but creative decisions still need a person.
  Source: https://www.reddit.com/r/podcasting/comments/1nurj29/is_ai_textbased_editing_the_future_of_video/

Interpretation:

• Assistant-stated but unverified: The practical market reality is modular: one app for recording and/or rough cutting, one for AI mini-clips, one for audio cleanup, and optionally one for Shorts.

3) Descript is still one of the strongest fits for talking-head rough cuts

• Verified: Descript pricing currently shows:
  • Hobbyist: $16/mobilledannually\ast \ast or\ast \ast$24/mo monthly
  • Creator: $24/mobilledannually\ast \ast or\ast \ast$35/mo monthly
  • Creator includes 4K export, full Underlord access, stock media, and “Generate video with the latest AI models.”
    Source: https://www.descript.com/pricing
• Verified (user-reported anecdote): Positive Reddit feedback commonly centers on text-based editing, captions, and transcript editing. Sources:
  • https://www.reddit.com/r/artificial/comments/1ntrxj5/out_of_all_the_ai_video_editing_tools_out_there/
  • https://www.reddit.com/r/podcasting/comments/1mtlvzx/anyone_else_here_using_descript_for_podcast/
• Verified (user-reported anecdote): Criticism also exists: interface clutter, occasional bugs, and some AI correction limitations on talking-head visual edits.
  Sources:
  • https://www.reddit.com/r/podcasting/comments/1mtlvzx/anyone_else_here_using_descript_for_podcast/
  • https://www.reddit.com/r/Descript/comments/17z5bkh/an_honest_review_of_descript_and_why_i_cancelled/

Conclusion:

• Verified + Assistant-stated but unverified: Descript is still one of the best first tools to trial for this exact use case, but it is not a perfect one-app production suite.

4) Riverside is strongest when recording quality, teleprompter, and local capture matter

• Verified: Riverside’s official pages confirm:
  • text-based editing
  • AI tools and Magic Clips
  • local recording
  • up to 4K video
  • separate tracks
  • up to 48 kHz audio
  • Pro plan at $24/moannually\ast \ast or\ast \ast$29/mo monthly
  • 15 hours of multi-track recording on Pro
    Sources:
  • https://riverside.com/
  • https://marketing.riverside.fm/pricing
• Verified (user-reported anecdote): Some users like Riverside’s end-to-end workflow and AI tools.
• Verified (user-reported anecdote): Some users complain about pricing changes and business-tier feature movement.
  Source: https://www.reddit.com/r/podcasting/comments/1ntksqg/riverside_pricing_bait_and_switch/

Conclusion:

• Assistant-stated but unverified: Riverside makes the most sense if Dmitri wants better recording flow, guest support, teleprompter, or remote interviews later.
• Assistant-stated but unverified: If Dmitri already has a workable recording method, Riverside is optional rather than mandatory.

5) Runway is one of the best-supported options for short AI-generated visual inserts

• Verified: Runway official pricing currently shows:
  • Standard: $12/user/month billed annually
  • Pro: $28/user/month billed annually
  • Standard includes 625 credits/month
  • Pro includes 2250 credits/month
    Source: https://runwayml.com/pricing
• Verified (user-reported anecdote): Current Reddit-style discussions continue to place Runway among the most-used options for fast AI clips, quick edits, and B-roll / filler shots, though these sources are anecdotal.
  Sources:
  • https://www.reddit.com/r/PartneredYoutube/comments/1ffr8bu/what_are_the_best_ai_tools_to_create_short_videos/
  • https://www.reddit.com/r/aitubers/comments/1r8ro27/what_b_roll_tools_do_you_use/
  • https://www.reddit.com/r/creativecoding/comments/1m7wv4u/tried_the_best_ai_texttovideo_tools_to_speed_up/

Conclusion:

• Assistant-stated but unverified: Runway is better treated as a mini visual insert tool than as a full long-form YouTube editor.

6) Pika is a credible lower-cost option for quick AI mini-clips

• Verified: Pika official pricing currently shows:
  • Basic: $8/month billed yearly
  • Standard: $28/month billed yearly
  • Pro: $76/month billed yearly
  • Basic includes 80 monthly video credits
  • Standard includes 700 monthly video credits
  • Pro includes 2300 monthly video credits
    Source: https://pika.art/pricing
• Verified (user-reported anecdote): Current discussions keep grouping Pika with Runway as one of the more common “usable clip out” options for quick AI visual generation.
  Sources:
  • https://www.reddit.com/r/PartneredYoutube/comments/1ffr8bu/what_are_the_best_ai_tools_to_create_short_videos/
  • https://www.reddit.com/r/aitoolforU/comments/1rw6pyl/what_ai_video_tool_actually_makes_content/

Conclusion:

• Assistant-stated but unverified: Pika is likely the best budget-first test bed for “AI B-roll / mini visual inserts” before spending more on Runway.

7) OpusClip is strong for long-to-shorts, but users still report review overhead

• Verified: OpusClip official pricing/features currently show:
  • Starter: $15/mo
  • Starter includes 150 credits/month
  • AI clipping, animated captions, auto-posting, filler/silence removal
  • Pro-level features include AI B-roll, social scheduler, export to Adobe Premiere Pro & DaVinci Resolve
    Source: https://www.opus.pro/pricing
• Verified (user-reported anecdote): Some users say OpusClip is decent or better than Descript/Riverside for auto-pulling shorts, but still requires review and manual fixes.
  Sources:
  • https://www.reddit.com/r/NewTubers/comments/1hmyp6l/had_you_ever_used_opusclip_or_any_other/
  • https://www.reddit.com/r/InstagramMarketing/comments/1qun0n6/ai_tool_for_creating_shorts_from_long_form_videos/

Conclusion:

• Assistant-stated but unverified: OpusClip is not the best first purchase unless Shorts repurposing is part of the channel strategy from day one.

8) CapCut is still popular for ease, but pricing and plan clarity are messy

• Verified: CapCut’s own resource page currently lists a Pro plan example at $19.99/month\ast \ast and\ast \ast$179.99/year.
  Source: https://www.capcut.com/resource/capcut-standard-vs-pro
• Verified (user-reported anecdote): Reddit users repeatedly report pricing inconsistency, promotion confusion, and jumps from legacy pricing to higher pricing.
  Sources:
  • https://www.reddit.com/r/CapCut/comments/1igso9o/difference_between_new_capcut_plans_free_vs/
  • https://www.reddit.com/r/CapCut/comments/1hntce0/why_did_capcut_increase_their_price_from_999_to/
  • https://www.reddit.com/r/CapCut/comments/1i8gz5f/price_increase_for_pro_renewal/
  • https://www.reddit.com/r/CapCut/comments/1iosm61/inconsistency_with_pro_pricing/

Conclusion:

• Assistant-stated but unverified: CapCut is still worth knowing, but it is a weaker “system of record” for a stable creator stack than Descript + one generator.

9) Audio cleanup is a hidden force multiplier

• Verified: Adobe Podcast Premium officially supports:
  • video support for Enhance Speech
  • bulk upload
  • enhancement strength controls
  • up to 4 hours/day
  • longer project/download limits in premium
  • current Adobe guide says $9.99/month\ast \ast or\ast \ast$99.99/year after trial
    Sources:
  • https://podcast.adobe.com/en/plans
  • https://podcast.adobe.com/en/guides/a-beginners-guide-to-text-based-editing
• Verified: Auphonic is officially free for 2 hours of processed audio per month and offers paid recurring or one-time credits beyond that.
  Source: https://auphonic.com/pricing

Conclusion:

• Assistant-stated but unverified: If budget is tight, audio cleanup is often a smarter spend than a second flashy AI video tool.

10) DaVinci Resolve remains an important free/one-time-cost fallback for final polish

• Verified: Blackmagic’s official pages confirm a free DaVinci Resolve version and a paid Studio version. The free version supports editing up to Ultra HD 3840x2160 at up to 60 fps.
  Sources:
  • https://www.blackmagicdesign.com/products/davinciresolve
  • https://www.blackmagicdesign.com/products/davinciresolve/studio
• Verified (user-reported anecdote): Reddit discussions continue to treat DaVinci as a credible low-cost serious editor, though with a steeper learning curve.
  Source: https://www.reddit.com/r/davinciresolve/comments/119ld7s/is_divinci_resolve_really_completely_free/

Conclusion:

• Assistant-stated but unverified: If Dmitri wants maximum polish later without adding another monthly subscription, DaVinci is the best upgrade path, but not the easiest first tool.

11) YouTube policy adds real constraints that were missing from the original answer

• Verified: YouTube requires creators to disclose content that is meaningfully altered or synthetically generated when it seems realistic.
  Source: https://support.google.com/youtube/answer/14328491
• Verified: YouTube explains “How this content was made” labels can appear when creators disclose altered or synthetic content or when YouTube’s own tools are used.
  Source: https://support.google.com/youtube/answer/15447836
• Verified: YouTube’s monetization policy states that reused content review checks whether the creator created, participated in, or produced the content.
  Source: https://support.google.com/youtube/answer/1311392
• Verified: YouTube’s 2024 and 2026 blog statements continue to emphasize disclosure of realistic synthetic media, while not requiring disclosure for clearly unrealistic/animated production assistance.
  Sources:
  • https://blog.youtube/news-and-events/disclosing-ai-generated-content/
  • https://blog.youtube/inside-youtube/the-future-of-youtube-2026/

Operational implication:

• Assistant-stated but unverified: Dmitri’s planned workflow is safer for authenticity and monetization than a faceless mass-generated template channel, because Dmitri is visibly participating, speaking, structuring, and producing the content.

Major Decisions and Conclusions

Recommended primary conclusion

• Verified + Assistant-stated but unverified: Dmitri should not attempt a fully AI-generated end-to-end workflow for the main channel.
• Verified + Assistant-stated but unverified: Dmitri should use real talking-head content as the core, with AI helping in the places where it saves the most pain:
  • transcript-based rough cuts,
  • filler/silence cleanup,
  • captions,
  • short AI-generated inserts,
  • audio enhancement,
  • optional Shorts repurposing.

Recommended stack choices

Best overall fit to start

• Assistant-stated but unverified: Descript Creator + Runway Standard + Adobe Podcast Premium
• Why:
  • Best balance of rough-cut speed, AI inserts, and audio polish
  • Affordable under the stated budget
  • Keeps Dmitri’s face/voice central
  • Avoids overreliance on synthetic full-video generation

Best ultra-lean test stack

• Assistant-stated but unverified: Descript Hobbyist + Pika Basic + Auphonic Free
• Why:
  • Cheapest serious test
  • Lets Dmitri validate process before committing more budget
  • Good enough for early pilots if the videos rely mostly on real footage

Best stack when recording quality/teleprompter matter

• Assistant-stated but unverified: Riverside Pro + Descript Hobbyist + Pika Basic
• Why:
  • Stronger recording flow
  • Better for remote guests later
  • Still comfortably under budget

Best stack if Shorts are strategically important

• Assistant-stated but unverified: Descript Creator + OpusClip Starter + Pika Basic
• Why:
  • Better for extracting repurposable clips
  • Works if each long-form video must also yield multiple shorts

Reasoning, Tradeoffs, and Why It Matters

Why transcript-first matters

• Verified (user-reported anecdote): Creators repeatedly describe the rough cut as the most painful part, and transcript-first editing reduces that pain.
  Source: https://www.reddit.com/r/youtubers/comments/1rqth58/i_finally_found_a_less_painful_way_to_rough_cut/

Why it matters:

• Assistant-stated but unverified: If the rough cut becomes easy, publishing consistency becomes much more realistic.

Why real footage should stay primary

• Verified: YouTube’s originality / reused-content logic rewards visible creator participation.
  Source: https://support.google.com/youtube/answer/1311392
• Verified (user-reported anecdote): Users consistently frame Runway/Pika style tools as better for inserts or quick clips than for full primary episodes.
  Sources:
  • https://www.reddit.com/r/PartneredYoutube/comments/1ffr8bu/what_are_the_best_ai_tools_to_create_short_videos/
  • https://www.reddit.com/r/aitubers/comments/1r8ro27/what_b_roll_tools_do_you_use/

Why it matters:

• Assistant-stated but unverified: This preserves authenticity, reduces artifact exposure, and lowers monetization / “reused content” risk.

Why audio is a bigger deal than many creators admit

• Verified: Adobe Podcast and Auphonic both explicitly center on fast voice cleanup and easier audio enhancement.
  Sources:
  • https://podcast.adobe.com/en/plans
  • https://auphonic.com/pricing

Why it matters:

• Assistant-stated but unverified: Better audio frequently makes a modest video feel professional, while weak audio makes even an expensive video feel amateur.

Why human review still cannot be removed

• Verified (user-reported anecdote): Users still report having to manually review Opus-generated clips and AI editing suggestions.
• Verified (user-reported anecdote): Users explicitly say AI is better at removing filler than at making creative story choices.
  Sources:
  • https://www.reddit.com/r/NewTubers/comments/1hmyp6l/had_you_ever_used_opusclip_or_any_other/
  • https://www.reddit.com/r/podcasting/comments/1nurj29/is_ai_textbased_editing_the_future_of_video/

Why it matters:

• Assistant-stated but unverified: The realistic win is not “no editing,” but “far less painful editing.”

Recommended Playbook / Process

Phase 1 — Prove the workflow cheaply

Recommended MVP stack

• Assistant-stated but unverified:
  1. Descript Hobbyist
  2. Pika Basic
  3. Auphonic Free
  4. Optional: existing phone/camera and existing mic

MVP goal

• Assistant-stated but unverified: Publish 3–5 videos before buying extra tools.
• Measure:
  • total time from idea to publish,
  • editing pain,
  • quality satisfaction,
  • need for more B-roll,
  • need for better recording flow,
  • whether Shorts repurposing is actually useful.

Phase 2 — Upgrade only after identifying the real bottleneck

Upgrade path if editing is still painful

• Assistant-stated but unverified: Move from Descript Hobbyist to Creator.

Upgrade path if AI visuals feel too weak or too limited

• Assistant-stated but unverified: Move from Pika Basic to Runway Standard or Pika Standard depending on which produces the style Dmitri actually likes.

Upgrade path if recording / teleprompter / guests become a bottleneck

• Assistant-stated but unverified: Add Riverside Pro.

Upgrade path if every video should create Shorts

• Assistant-stated but unverified: Add OpusClip Starter.

Per-video workflow

1) Script / outline

• Assistant-stated but unverified: Use AI to draft:
  • hook,
  • 3–5 main points,
  • transitions,
  • CTA,
  • notes for where visuals are needed.
• Assistant-stated but unverified: Put visual prompts directly in the outline using tags like:
  • [visual: abstract network nodes forming]
  • [visual: secure systems dashboard]
  • [visual: paperwork chaos, then clean workflow]

2) Record Dmitri’s talking-head footage

• Assistant-stated but unverified: Record the main explanation as the A-roll backbone.
• Assistant-stated but unverified: Keep the delivery intentional:
  • pause between sections,
  • repeat lines cleanly when needed,
  • leave room for later cutaways.

3) Rough cut via transcript

• Verified: Descript and Riverside both support text-based editing.
  Sources:
  • https://www.descript.com/pricing
  • https://riverside.com/
• Assistant-stated but unverified: Do not polish visuals yet. First:
  • remove obvious tangents,
  • remove repeated takes,
  • tighten pacing,
  • cut filler,
  • verify structure.

4) Clean audio

• Verified: Adobe Podcast Premium and Auphonic both offer fast audio enhancement.
  Sources:
  • https://podcast.adobe.com/en/plans
  • https://auphonic.com/pricing
• Assistant-stated but unverified: Run this before final export so every visual choice sits on top of strong audio.

5) Add AI-generated mini-clips sparingly

• Verified (user-reported anecdote): AI-generated B-roll is commonly used in short bursts to keep artifacts less noticeable.
  Sources:
  • https://www.reddit.com/r/AI_UGC_Marketing/comments/1qgfb6k/what_is_broll_in_ai_video_can_ai_generate_broll/
  • https://www.reddit.com/r/aitubers/comments/1r8ro27/what_b_roll_tools_do_you_use/
• Assistant-stated but unverified: Best practice:
  • keep most AI clips around 2–5 seconds
  • use them for emphasis, transitions, and concept illustration
  • avoid leaning on them for long sustained scenes
  • do not let AI B-roll replace the core human narrative

6) Add captions and export

• Verified: Descript, Riverside, OpusClip, Adobe Podcast, and CapCut all position captions as part of their workflows in current materials or user reports.
  Sources: official product pages above plus relevant Reddit references.
• Assistant-stated but unverified: Use captions selectively. For long-form YouTube, lighter captions often look more premium than hyperactive short-form styling.

7) Repurpose to Shorts if that is actually part of strategy

• Verified: OpusClip is explicitly built for clipping, captions, and social publishing.
  Source: https://www.opus.pro/pricing
• Assistant-stated but unverified: Only add this step if it serves the channel. Do not create a second workflow just because the tool exists.

Practical style rules for a more professional result

• Assistant-stated but unverified: Use AI visuals as support, not spectacle.
• Assistant-stated but unverified: Let the viewer see Dmitri early and often; do not hide the human source of expertise.
• Assistant-stated but unverified: Prefer fewer, cleaner transitions over constant movement.
• Assistant-stated but unverified: Use a consistent frame, background, and lighting so the channel looks intentional.
• Assistant-stated but unverified: Prioritize:
  1. understandable structure,
  2. clean audio,
  3. decent lighting,
  4. clean captions,
  5. only then more AI visuals.

Budget Stack Options

Stack	Components	Estimated recurring software cost	Status
Lean test stack	Descript Hobbyist + Pika Basic + Auphonic Free	~$24/mo before tax	Assistant-stated but unverified budget assembly using verified prices
Balanced stack	Descript Creator + Runway Standard + Adobe Podcast Premium	~$45.99/mo before tax	Assistant-stated but unverified budget assembly using verified prices
Recording-focused stack	Riverside Pro + Descript Hobbyist + Pika Basic	~$48/mo before tax	Assistant-stated but unverified budget assembly using verified prices
Long + Shorts stack	Descript Creator + OpusClip Starter + Pika Basic	~$47/mo before tax	Assistant-stated but unverified budget assembly using verified prices
Premium but under cap	Riverside Pro + Descript Creator + Runway Standard + Adobe Podcast Premium + OpusClip Starter	~$84.99/mo before tax	Assistant-stated but unverified budget assembly using verified prices

Budget caveats

• Verified: Several of these prices rely on annual billing discounts.
• Tentative / speculative: Taxes, regional pricing, promo pricing, and changing credits may materially alter total cost.
• Tentative / speculative: Generative video credit burn can exceed expectations quickly if prompts require many retries.

Tools, Resources, Links, and References

Tool decision matrix

Tool	Primary role	Current pricing / plan notes	Best use	Main concerns	Evidence status
Descript	Transcript-first editing, captions, cleanup, AI assist	Hobbyist $16annual/$24 monthly; Creator $24annual/$35 monthly	Talking-head rough cuts	UI clutter, occasional bugs, not a total replacement for judgment	Verified
Riverside	Recording + text editing + repurposing	Standard $15annual/$19 monthly; Pro $24annual/$29 monthly	Better recording flow, teleprompter, guests, local capture	Some pricing / business-plan complaints	Verified
Runway	AI video inserts / generation	Standard $12annual;Pro$28 annual	Short AI mini-clips and support visuals	Credit use, experimentation overhead	Verified
Pika	Lower-cost AI video inserts	Basic $8annual;Standard$28 annual; Pro $76 annual	Budget AI cutaways	Quality/consistency depends on prompts	Verified
OpusClip	Long-to-shorts clipping	Starter $15/mo; Pro/custom tiers	Shorts extraction	Still needs review	Verified
Adobe Podcast	Audio cleanup, browser editing	Official guide states $9.99/moor$99.99/year	Easy audio polish	Not a full video stack	Verified
Auphonic	Audio post-processing	Free for 2 hours/month, then credits/plans	Cheap audio cleanup	Limited free allowance	Verified
CapCut	Easy visual editor	Official resource page example shows Pro $19.99/mo,$179.99/year	Fast editing/templates	Pricing ambiguity and plan churn	Verified with caveat
DaVinci Resolve	Final polish / serious edit	Free version available; Studio paid	Final polish without ongoing sub	Harder to learn	Verified

Source list

Official / primary sources

1. Descript pricing: https://www.descript.com/pricing
2. Riverside homepage: https://riverside.com/
3. Riverside pricing: https://marketing.riverside.fm/pricing
4. Runway pricing: https://runwayml.com/pricing
5. Pika pricing: https://pika.art/pricing
6. OpusClip pricing: https://www.opus.pro/pricing
7. Adobe Podcast plans: https://podcast.adobe.com/en/plans
8. Adobe Podcast guide with price statement: https://podcast.adobe.com/en/guides/a-beginners-guide-to-text-based-editing
9. Auphonic pricing: https://auphonic.com/pricing
10. CapCut pricing/resource page: https://www.capcut.com/resource/capcut-standard-vs-pro
11. DaVinci Resolve official product page: https://www.blackmagicdesign.com/products/davinciresolve
12. DaVinci Resolve Studio page: https://www.blackmagicdesign.com/products/davinciresolve/studio
13. YouTube altered/synthetic disclosure policy: https://support.google.com/youtube/answer/14328491
14. YouTube “How this content was made” labels: https://support.google.com/youtube/answer/15447836
15. YouTube monetization / reused content policy: https://support.google.com/youtube/answer/1311392
16. YouTube 2024 disclosure announcement: https://blog.youtube/news-and-events/disclosing-ai-generated-content/
17. YouTube 2026 outlook / disclosure note: https://blog.youtube/inside-youtube/the-future-of-youtube-2026/

User-reported / Reddit / anecdotal sources

18. Transcript-first rough cuts feel less painful: https://www.reddit.com/r/youtubers/comments/1rqth58/i_finally_found_a_less_painful_way_to_rough_cut/
19. AI text editing still needs human creative judgment: https://www.reddit.com/r/podcasting/comments/1nurj29/is_ai_textbased_editing_the_future_of_video/
20. Record in Riverside, edit in Descript, clip in Opus: https://www.reddit.com/r/podcasting/comments/1redo6x/is_riverside_business_worth_it_for_webinars/
21. OpusClip still requires review: https://www.reddit.com/r/NewTubers/comments/1hmyp6l/had_you_ever_used_opusclip_or_any_other/
22. Descript strengths for transcript editing and AI video assistance: https://www.reddit.com/r/artificial/comments/1ntrxj5/out_of_all_the_ai_video_editing_tools_out_there/
23. Descript criticism / clutter / moving parts: https://www.reddit.com/r/podcasting/comments/1mtlvzx/anyone_else_here_using_descript_for_podcast/
24. Older Descript criticism around talking-head corrections: https://www.reddit.com/r/Descript/comments/17z5bkh/an_honest_review_of_descript_and_why_i_cancelled/
25. Riverside pricing complaint / business-tier frustration: https://www.reddit.com/r/podcasting/comments/1ntksqg/riverside_pricing_bait_and_switch/
26. Pika / Runway / quick AI clips for B-roll: https://www.reddit.com/r/PartneredYoutube/comments/1ffr8bu/what_are_the_best_ai_tools_to_create_short_videos/
27. B-roll tools discussion including Runway, Pika, Opus: https://www.reddit.com/r/aitubers/comments/1r8ro27/what_b_roll_tools_do_you_use/
28. AI-generated B-roll clip length suggestion: https://www.reddit.com/r/AI_UGC_Marketing/comments/1qgfb6k/what_is_broll_in_ai_video_can_ai_generate_broll/
29. CapCut pricing confusion thread: https://www.reddit.com/r/CapCut/comments/1igso9o/difference_between_new_capcut_plans_free_vs/
30. CapCut increase complaint: https://www.reddit.com/r/CapCut/comments/1hntce0/why_did_capcut_increase_their_price_from_999_to/
31. CapCut renewal complaint: https://www.reddit.com/r/CapCut/comments/1i8gz5f/price_increase_for_pro_renewal/
32. CapCut inconsistency thread: https://www.reddit.com/r/CapCut/comments/1iosm61/inconsistency_with_pro_pricing/

Risks, Caveats, and Red Flags

1) Pricing volatility

• Verified: CapCut shows clear plan/pricing confusion in user reports, and Riverside also has some complaints around plan changes.
• Tentative / speculative: Other tools can also change pricing, credits, and feature boundaries with little notice.

2) Credit burn on AI video tools

• Verified: Runway and Pika both use credit-based systems.
• Assistant-stated but unverified: AI video generation often costs more in practice than it appears, because retries, prompt refinement, and style matching consume credits fast.

3) Human review is not optional

• Verified (user-reported anecdote): Users still review generated clips and edit results.
• Assistant-stated but unverified: Skipping review is the fastest way to publish awkward, generic, or artifact-ridden content.

4) AI visuals can quickly look cheap

• Verified (user-reported anecdote): User reports consistently frame AI visuals as strongest for short inserts, not full episodes.
• Assistant-stated but unverified: Overusing them will make the channel feel synthetic rather than premium.

5) Tool overlap can create waste

• Verified: Descript, Riverside, OpusClip, CapCut, and Adobe all overlap around editing/captions/AI features.
• Assistant-stated but unverified: Buying several overlapping subscriptions too early is a common mistake.

6) YouTube disclosure and originality issues

• Verified: Realistic synthetic content may require disclosure.
• Verified: Monetization review checks originality and creator participation.
• Assistant-stated but unverified: Dmitri’s workflow is safer than faceless template automation, but bad execution could still make videos look mass-produced.

7) AI full-video promises are often overstated

• Verified (user-reported anecdote): People still combine tools and still talk about manual tweaks.
• Assistant-stated but unverified: Any tool claiming true one-click, channel-quality long-form production should be treated skeptically.

Open Questions / What Still Needs Verification

1. Tentative / speculative: Which AI video generator Dmitri personally likes better for Solanasis aesthetics: Runway vs Pika vs possibly another newer entrant (for example, Kling, Luma, or Adobe Firefly video tools) still needs hands-on testing.
2. Tentative / speculative: Exact monthly spend in practice depends on how many retries Dmitri needs in Runway/Pika, which cannot be verified from pricing pages alone.
3. Tentative / speculative: CapCut’s real current price for Dmitri may differ from the official resource page because user reports suggest storefront/device/region promo inconsistency.
4. Tentative / speculative: Whether Riverside adds enough value over a phone/camera + Descript depends on Dmitri’s actual recording setup and whether teleprompter / guests are near-term needs.
5. Tentative / speculative: Whether OpusClip meaningfully saves time for Dmitri versus manual Shorts editing depends on content style and tolerance for cleanup.
6. Assistant-stated but unverified: The exact “best-looking” workflow for Dmitri’s brand probably depends more on:
   • lighting,
   • framing,
   • audio,
   • pacing,
   • restrained B-roll use, than on the specific AI generator.
7. Tentative / speculative: If Dmitri wants an even more polished result later, a hybrid workflow using DaVinci Resolve for final polish may outperform an all-browser tool stack, but that adds skill overhead.

Suggested Next Steps

Immediate next step

• Assistant-stated but unverified: Pilot this exact stack for 3 videos:
  1. Descript Hobbyist or Creator
  2. Pika Basic or Runway Standard
  3. Adobe Podcast Premium or Auphonic
• Success criteria:
  • publishable long-form result,
  • acceptable quality,
  • workflow not miserable,
  • software spend still low.

Next validation step

• Assistant-stated but unverified: Compare two pilot workflows:
  • Workflow A: Descript + Pika + audio cleanup
  • Workflow B: Riverside + Descript + Pika + audio cleanup
    Determine whether Riverside actually removes enough friction to justify itself.

Next systems step

• Assistant-stated but unverified: Build a repeatable SOP with:
  • script template,
  • visual prompt library,
  • talking-head shot checklist,
  • caption style rules,
  • export checklist,
  • YouTube upload/disclosure checklist,
  • file naming convention,
  • asset folder structure.

Strategic next step

• Assistant-stated but unverified: Decide whether Shorts are a core part of the channel strategy. If yes, evaluate OpusClip. If not, do not add it yet.

Handoff Notes for Another AI

1. The user does not want a faceless spammy AI channel. Keep Dmitri visibly involved as the speaker and expert.
2. The main optimization target is editing pain reduction, not maximum AI spectacle.
3. The budget target is under $100/month, ideally materially lower for the first phase.
4. The strongest current recommendation is still:
   • Descript for rough cut
   • Runway or Pika for short AI visual inserts
   • Adobe Podcast or Auphonic for audio cleanup
5. A modular stack is more credible than a one-app stack.
6. Maintain caution around:
   • YouTube disclosure rules,
   • originality / reused-content monetization risks,
   • pricing churn,
   • credit burn,
   • human review requirements.
7. If continuing this work, the best next deliverables would be:
   • a Solanasis YouTube production SOP
   • a prompt pack for AI B-roll / mini-clips
   • a weekly publishing workflow
   • a recording setup guide
   • and a tool trial matrix for 3–5 pilot videos.

Reviewer Notes and Improvements Made

No reviewer-agent capability was available in this environment, so a serious self-review pass was performed.

Improvements made over the original answer

• Added explicit evidence-status labeling for every major point.
• Re-verified current official pricing/features where possible.
• Preserved official source links and user-reported Reddit links separately.
• Added YouTube policy constraints that were missing from the original answer.
• Converted the prior narrative answer into a decision memo + implementation playbook.
• Added risk analysis, open questions, and a handoff section for another AI.
• Reduced unsupported certainty around anecdotal Reddit claims.
• Clarified that “Verified” Reddit items mean the report exists, not that the claim is universally true.
• Flagged areas where hands-on testing is still the only reliable way to decide.

Optional Appendix with structured YAML-style summary

document_type: research-grade handoff memo and playbook
date: 2026-03-18
user_goal:
  - create professional YouTube videos
  - Dmitri still talks on camera
  - use AI for editing relief, captions, mini-clips, cleanup
  - stay under roughly 100 USD per month
core_conclusion:
  - use real talking-head footage as the backbone
  - use transcript-based editing for rough cuts
  - use AI mini-clips/B-roll selectively
  - use audio cleanup
  - optionally add long-to-shorts repurposing
primary_recommendation:
  stack:
    - Descript Creator
    - Runway Standard
    - Adobe Podcast Premium
  estimated_cost_before_tax: 45.99
lean_test_stack:
  - Descript Hobbyist
  - Pika Basic
  - Auphonic Free
  estimated_cost_before_tax: 24.00
recording_upgrade_stack:
  - Riverside Pro
  - Descript Hobbyist
  - Pika Basic
  estimated_cost_before_tax: 48.00
shorts_stack:
  - Descript Creator
  - OpusClip Starter
  - Pika Basic
  estimated_cost_before_tax: 47.00
major_risks:
  - pricing churn
  - credit burn
  - AI visuals looking cheap if overused
  - human review still required
  - YouTube disclosure/originality constraints
best_next_deliverables:
  - solanasis_youtube_production_sop
  - ai_broll_prompt_library
  - weekly_publishing_workflow
  - recording_setup_checklist
  - pilot_tool_trial_matrix