Pay Only For What You Use

Simple, transparent pricing. No subscriptions, no commitments. Add credits and use them whenever you need.

$0.15 1080p Video newtiska'q second · $0.015 Still

Credits Ika'tu

What's Free

Free every day from your first top-up: 20 AI design turns on a free model and 5 previews (a still, a widget or a short GIF) when your balance cannot cover one, from a daily pool shared by everyone and refreshed at midnight UTC. The Studio and the open-source MCP server are free on your own device, no account needed.

$10
Minimum top-up
$0.07
A typical AI design turn
No
Expiration

Credits can't be bought in the app during the beta.

Add Credits

Buy credits once, use them anytime. Credits never expire. You receive the full amount minus Stripe processing fees.

$10
$9.41 in credits
≈ 62 Video (newtiska'q second) · 627 Still · kisna ~135 AI design Kisi'tu
Get Started
$50
$48.25 in credits
≈ 321 Video (newtiska'q second) · 3,216 Still · kisna ~696 AI design Kisi'tu
Credits Ika'tu
$100
$96.80 in credits
≈ 645 Video (newtiska'q second) · 6,453 Still · kisna ~1,397 AI design Kisi'tu
Credits Ika'tu

Or buy any amount from $10 to $1,000 at /add-credits — the same 2.9% + $0.30 card fee comes off every purchase.

What You Pay For

PinePaper Studio is free to use. You create art in your browser and download it locally at no cost. Cloud credits pay for hosted work: MCP tool calls, cloud renders, voiceovers, the $0.002 platform fee on voice and sound made with your own vendor key, and events collected by published widgets ($0.01 per 1,000, billed to the widget's owner).

1. MCP Tool Requests

When an AI assistant (Claude Desktop, or our chat) creates graphics via MCP protocol, each tool call costs credits. This covers Cloudflare Workers compute.

MCP tool call (create, animate, query) $0.001
Background generator $0.002
Template save(size-based) $0.001+
Conversation sync(size-based) $0.001+

2. Cloud Exports

Local export is free but depends on your hardware — complex scenes with many items can be slow or fail on lower-end devices. Cloud rendering handles any complexity consistently, and is the only option for mobile devices or cloud-only sessions.

Renders are priced by frames, not by a flat rate — a 720p clip is genuinely half the work of 1080p, so it costs half. One credit is $0.015, and buys one second of 1080p/30 MP4.

credits = ceil( seconds × fps/30 × resolution × format )
Resolution720p · 1080p · 4K ×0.5 · ×1 · ×4
FormatMP4/WebM · GIF · animated SVG ×1 · ×1.5 · ×0.25
PNG or PDFa single frame, any resolution $0.015
Embeddable widget or editable projectone build, any resolution $0.015

Any whole frame rate from 1 to 120 fps. Up to 600 seconds per render job; a longer piece renders in segments, up to 3,600 seconds in total. Factors and a calculator: /render-pricing.

What that comes to:

10s MP4 · 1080p · 30fps $0.15
10s MP4 · 720p · 30fps $0.075
10s GIF · 1080p · 30fps $0.225
10s animated SVG · 1080p $0.045

POST /v1/renders:estimate returns the exact credits and price for any profile, free, before you spend anything.

Pipelines: hundreds of variants from a feed

A feed rendering around the clock pays the same formula as one render from the Studio — the price per credit does not change with volume.

500 variants · 6s MP4 · 1080p3,000 credits $45
500 variants · 6s MP4 · 720p1,500 credits $22.50
Re-running the feedrows that did not change come back cached $0

Quote the whole feed first, free: POST /v1/renders:batchEstimate prices up to 500 rows in one call, each exactly as :estimate would. It is authoritative — if this page and the endpoint ever disagree, the endpoint is right. POST /v1/renders:batchCheck (also free) says, field by field, which rows of a template feed would overflow, run off the frame or miss a font, before anything is reserved. A pipeline runs on a service-account key, which can carry its own spend cap; the daily spending cap for people does not apply to it. How it fits together: /docs/agents.

3. Voice, Sound & Embeds

Platform voiceovers follow Google's list rate on the day and are charged for the seconds actually spoken — never more than the free quote from POST /v1/voice:estimate.

Voiceover · flashthrough 31 Dec 2026 · about $0.08/min from 1 Jan 2027 ~$0.04/min
Voiceover · litethrough 31 Dec 2026 · about $0.05/min from 1 Jan 2027 ~$0.03/min
Voice on your own keyElevenLabs, OpenAI or Cartesia · the vendor bills your key $0.002
Generated sound effecton your ElevenLabs key · the vendor bills your key $0.002
Generated music or sound effectStable Audio 2.5 on our key · up to 190 s · nothing charged if it fails $0.60
Sound librarylicensed music and effects · private beta · render and download, not yet for publishing Free in beta
A film's own scorePinePaper's built-in instruments, mixed into your export Included
Embed eventscollected by a published widget, billed to its owner $0.01 / 1,000

On your own key, the audio's licence follows your plan with that vendor. Transcripts, image edits and uploads are currently free; keeping files is covered by the storage note below.

LLM Chat (Optional)

The built-in chat lets you describe what to create in natural language. You choose how to power it:

Option LLM Cost
PinePaper LLM COMING SOON All-in-one: LLM + style learning + knowledge graph — billed from credits when it launches $0.20/1M input · $0.60/1M output tokens
Hosted models Claude, Gemini and open models — no key needed. Billed from your credits by the tokens a turn uses; each model's rate is shown in the chat before you send By tokens
Bring Your Own Key (BYOK) Use your own Claude, Gemini, GPT, or Grok key $0.002 platform fee + provider costs
pinepaper-llm-v0 our own design model · Qwen3.5-9B fine-tune COMING SOON

An intent-aware model purpose-built for PinePaper: it will classify your instruction (design vs. chat), emit canvas tool calls directly, stream token-by-token, and learn your style over time via the private knowledge graph. Built by fine-tuning a compact ~9B open base (Qwen3.5-9B, Apache 2.0) and served with vLLM — small enough to keep per-token cost low. Not yet available. Until it is, the chat runs on the hosted models above, billed from your credits, or on your own API key (BYOK).

Input $0.20 / 1M tokens Output $0.60 / 1M tokens Context 32K tokens Max output 12K tokens Streaming yes

Planned pricing. A typical design turn (~1.5K in / 0.5K out) ≈ $0.0006 — about 1,500 turns per $1 of credits once it's live.

BYOK: You pay the provider directly at their rates. We charge $0.002 per request for infrastructure (Cloudflare Workers, KV storage, session management). See provider pricing: Anthropic • Google AI • OpenAI • xAI

Storage: Templates, conversations and uploads you keep with us: the first 100 MB are free, then $1.00 per GB a month. This is not charged yet — your account page shows what it would be, and you will see it there before any charge starts.

Don't need AI chat? You can use PinePaper Studio without it. Create graphics manually, use templates, and export locally — all free.

Real Cost Examples

100 MCP tool calls

$0.10

AI design turn, hosted model

~$0.069 a turn

measured on 481 real design turns, 2026-10-10 to 2026-10-16: Gemini $0.088 · Claude $0.019 · Open model (Cerebras) $0.010. Each model's rate is shown in the chat before you send.

100 BYOK chats

$0.20 platform fee

+ your provider costs

Local export

Free

Download from Studio

Agencies, designers and teams

Building something bigger than one person's account? Write to us at support@pinepaper.studio and say what you need. Below is what exists today; for anything else, we will tell you honestly whether we can do it.

PinePaper is made by one developer for now, so there is no 24/7 support or SLA. If something is not working, write to support@pinepaper.studio: every message is read, and answered as soon as we can.

Custom work

A template built for you

Need a template in your brand that the catalogue does not have? Tell us what it should say and show, and we will reply with whether and how we can make it.

Ask about a template
Designers & agencies

Work for clients

Make films and templates for the people you work for, and send them to the page or the file they need.

  • Sign-ups through your handle's link are counted (a count only; no rewards): ?ref=your_handle
  • Your name, licence and mark on every render, set once on your creator profile
  • Public preview links for any finished render
Talk to us
Pipelines

Rendering at volume

For feeds and pipelines that render many films from one template.

  • Service-account keys, each with its own spend cap
  • Every render priced free before it is made
  • Signed completion webhooks and batch reconciliation
Ask about volume

How It Works

1

Add Credits

Pay once with any amount ($10 minimum). Get your API key instantly.

2

Use Your API Key

Configure Claude Desktop or any MCP client with your API key.

3

Create & Export

Use credits for MCP requests and exports. Credits never expire.

Frequently Asked Questions

Why is there a $10 minimum?

Stripe charges $0.30 + 2.9% per transaction. With a $10 minimum, you get $9.41 in credits - a reasonable value. Smaller amounts would lose too much to fees.

Do credits expire?

No! Your credits never expire. Add $10 today and use them over the next year if you want. They're yours until you use them.

Can I get a refund?

Unused paid credits can be refunded within 7 days of purchase. The refund equals your remaining unused paid credit balance (promo credits excluded). After 7 days, all purchases are final — but credits never expire. See our Terms of Service for full details.

What is free?

Free every day from your first top-up: 20 AI design turns on a free model and 5 previews (a still, a widget or a short GIF) when your balance cannot cover one, from a daily pool shared by everyone and refreshed at midnight UTC. The Studio and the open-source MCP server are free on your own device, no account needed.

What does $1 buy?

66 stills (PNG or PDF), 6 ten-second 1080p films, or 22 ten-second animated SVGs — each priced by the same formula as POST /v1/renders:estimate.

Do I need credits to export?

Basic exports are free locally. But complex scenes with many canvas items can be slow or crash on weaker hardware. Cloud export renders on our servers — consistent quality regardless of scene complexity. Also the only option on mobile or cloud-only sessions.

Can I use the free open-source version?

Yes. PinePaper Studio is free to use in your browser. The MCP Server is MIT licensed — run it locally with Claude Desktop for free AI-assisted creation. Cloud credits pay for hosted work: MCP tool calls, cloud renders, voiceovers, and the small platform fees listed on this page.

How do I check my credit balance?

Sign in and open your dashboard at /account — it shows your balance, what you have spent and your recent requests. Programmatically, GET /api/me/account with your session.

What is BYOK mode?

Bring Your Own Key (BYOK) lets you use your existing API keys from Claude, GPT, Gemini, or Grok. You pay the provider directly with no LLM markup. A $0.002/request platform fee covers our infrastructure (KG pipeline, sessions, caching).

Do you offer design services or agency partnerships?

Not as a fixed offer yet. Write to support@pinepaper.studio with what you need, a template built in your brand or work for clients, and we will tell you honestly whether we can do it. Referral links (?ref=your_handle) are attributed today; there is no revenue share programme.