John Nordström

AI director & filmmaker. Festival Official Selection ×2. Building cinematic work with AI tools — and GenSpend, the spend tracker for people who actually use the

[John Nordström @synthden]

If you make things with AI tools, you know the problem: Runway, Kling, Midjourney, ElevenLabs — and no real picture of what it costs.

I built GenSpend. One ledger across your AI tools, credits converted to real dollars, true cost per finished piece.

https://t.co/SaL01FQIBO

[John Nordström @synthden]

@AbuMusa3b Great work! Quick question: when you price AI client work, how are you tracking credits, retries, and subscriptions across one deliverable?

[John Nordström @synthden]

@ze_bigfoot Quick question: when you price AI client work, how are you tracking credits, retries, and subscriptions across one deliverable?

[John Nordström @synthden]

@Turboslowagency @anilaltas Quick question: when you price AI client work, how are you tracking credits, retries, and subscriptions across one deliverable?

[John Nordström @synthden]

@Genflickmovies Quick question: when you price AI client work, how are you tracking credits, retries, and subscriptions across one deliverable?

[John Nordström @synthden]

@alimamakx The hard part is not knowing your monthly AI spend. It is knowing what this specific client deliverable actually cost.

[John Nordström @synthden]

@Mavernxbt Curious: are you tracking this per client/project now, or just looking at tool subscriptions after the fact?

[John Nordström @synthden]

@NutopiaAI @xai Ha — right? Once you start counting keepers instead of gens, the whole "which is cheaper" question flips. Good thread, man.

[John Nordström @synthden]

@dustinhollywood @capcutapp @dreamina_ai @stages_ai @imagine @Kling_ai @Hailuo_AI @midjourney @reve ECLIPTIC looks great — serious stack. Curious how Stages handles spend outside its own walls though? CapCut, MJ, Kling, Hailuo all encode credits differently, so one real-dollar number per deliverable across all of them is the part that drove me nuts. Appreciate the follow, man.

[John Nordström @synthden]

@vschmalbach Yeah, that's exactly the part that's deceptively annoying — every vendor encodes credits differently, so the real work is normalizing it all into actual dollars per deliverable. That's what GenSpend does. It's live if you ever want to poke at it: https://t.co/SaL01FQIBO

[John Nordström @synthden]

If you make things with AI tools, you know the problem: Runway, Kling, Midjourney, ElevenLabs — and no real picture of what it costs.

I built GenSpend. One ledger across your AI tools, credits converted to real dollars, true cost per finished piece.

https://t.co/SaL01FQIBO

[John Nordström @synthden]

A client AI video might touch Runway, Kling, ElevenLabs, Midjourney, Topaz, and a few retries nobody writes down. Vendor dashboards show tool usage. They do not answer: what did this client deliverable cost me?

[John Nordström @synthden]

@NutopiaAI @xai 14/day is rough when half are unusable — the messy ones still burn the gen. Seedance hit rate is way better per credit, even if the sticker looks higher. Quality-per-dollar beats gens-per-day

[John Nordström @synthden]

@allendpresents @madgreek65 Yeah — the mixed stack is smart but it's exactly why the real per-deliverable cost gets invisible. Free Gemini for some, paid for others, OSS for the rest... no single number tells you what a finished piece actually cost.

[John Nordström @synthden]

@madgreek65 @allendpresents The 5–20 min ones are sneaky — feels cheap per gen, then the month's bill lands and it's the small stuff that added up. OSS move makes sense if your volume's steady

[John Nordström @synthden]

@AvelyrahnAI How are you tracking your credit spend with the different AI tools?

[John Nordström @synthden]

@HodlReaper How are you tracking your credit spend with the different AI tools?

[John Nordström @synthden]

@rugikkk How are you tracking your credit spend with the different AI tools?

[John Nordström @synthden]

@georgesttock The hard part is not knowing your monthly AI spend. It is knowing what this specific client deliverable actually cost.

[John Nordström @synthden]

@georgesttock Claude

[John Nordström @synthden]

@Framer_X The hard part is not knowing your monthly AI spend. It is knowing what this specific client deliverable actually cost.