Billing & credits

anymd bills in credits, and you pay when your agents learn something new. Processing a new source uses credits by its complexity (one credit is one web page); reusing what your agents already know is free: cached reads, library search and MCP recall cost nothing.

What things cost

Action	Credits
Web page, GitHub, Reddit, Hacker News, X post	1
Facebook, Instagram, Threads or LinkedIn post	10
Each additional X thread post	1
Each started batch of 20 comments or replies	10
Each successfully analyzed article image	5
YouTube video with transcript	3
YouTube video stored on the anymd CDN (opt-in setting, charged when ready)	20
AI analysis of a stored YouTube video (opt-in setting, charged when ready)	10 + 20 per started minute
PDF, DOCX, XLSX, CSV and other documents (per file)	3
Image (vision description)	5
Social search: each page of results on X, Facebook, Instagram or Threads	10
Social search: each page of LinkedIn results	100
Social search on all platforms	Sum of the platforms that returned results
Social search page with no results	0
Library search, reads, MCP reads	0
Cached result	0
Failed conversion	0

Every conversion response tells you what it cost in the X-Anymd-Credits header. fresh=1 skips the cache, so it is charged like a new conversion.

Deep reading

Thread expansion, comments and article-image analysis can spend credits beyond the base price, so they are off by default and run only when you opt in: on a single request, or as a saved default in Account → Deep reading defaults (also PUT /api/v1/account/reading-preferences and anymd prefs set). Signing in never turns them on. One rule decides each option: request option > your saved default > safe default (off). A one-time change in the converter or on a request does not change your saved defaults.

Option	Default	Bounds	Credits
Keep image/media URLs (images, keepImages)	On		0
Expand X threads (expandThread)	Off	maxThreadPosts 1–100, default 20	1 per additional post
Comments & replies (includeComments)	Off	maxComments 1–1,000, default 100	10 per started batch of 20
Read text & details in images (analyzeImages)	Off	maxImages 1–20, default 10	5 per analyzed image
Max credits per conversion (maxCredits)	100	1–1,000	Cap including the base price
Download YouTube videos to the anymd CDN (downloadVideo)	Off	Per request or saved; account required	20 per video, charged when the download is ready
Analyze downloaded videos with AI (analyzeVideo)	Off	Per request (turns on downloadVideo) or saved (needs downloadVideo); videos up to 60 minutes	10 + 20 per started minute of video, charged when the analysis is ready

downloadVideo and analyzeVideo can be sent with each request or saved as defaults, and run outside the conversion: they are not part of maxCredits or credit_breakdown. Each finished download is a separate video_download row in your usage log, and each finished analysis a video_analysis row. A failed download or analysis costs nothing, and reading the same video again while its job is running or ready reuses that job instead of charging twice.

The worst case is bounded before you convert: for example, threads up to 20 posts add at most 19 credits, and 10 analyzed images add at most 50, always capped by maxCredits. Saved defaults never raise your plan's allowance, and a request's own maxCredits wins over the saved one. Each response's credit_breakdown (base, thread, comments, images) and the trace in Dashboard → Traces show which enrichment produced a charge and whether it came from the request, your saved default or the safe default.

Until October 2026, signed-in X conversions expanded same-author threads automatically. They now return the single requested post unless thread expansion is enabled.

maxCredits defaults to 100 and accepts 1–1,000; maxThreadPosts defaults to 20 and accepts 1–100; maxComments defaults to 100 and accepts 1–1,000; maxImages defaults to 10 and accepts 1–20. A request is bounded at 40 provider calls and 55 seconds. Successful returned units consume credits; failed or unavailable units do not. Partial sections are reported in the response's enrichment coverage instead of being charged as complete.

Plans

Plan	Price	Credits / month	Beyond included credits
Free	$0	500	Hard stop until next month
Pro	$9/mo, or $7/mo billed yearly	10,000	$1 per extra 1,000
Scale	$49/mo, or $39/mo billed yearly	100,000	$0.60 per extra 1,000
Enterprise	Custom	Custom volume	$0.40 per 1,000

The pricing page is always current and lists each plan's extras (library size, key limits, support).

• Free keeps working for search and reads after credits run out; new conversions return 402 quota_exceeded until the next month.
• Pro and Scale never block: usage beyond the included credits is metered and billed as overage. Credits granted by support are used after the included credits and before overage, from the moment they are granted: they keep later usage off the overage bill but do not refund overage already billed.
• Credits reset at the start of each calendar month (UTC).

No account?

Anonymous use of the URL API is free up to 50 conversions per day per IP. Nothing is saved.

Offers

Code	Discount	Who and when
LAUNCH30	30% off	Launch offer, until 2026-10-31
COMEBACK20	20% off	Returning visitors, valid for 48 hours after it's shown

Enter the code at checkout.

Checkout and managing your plan

Payments are handled by Polar.sh, our merchant of record. Polar collects the payment and handles sales tax and VAT.

• Upgrade: from the pricing page or your dashboard, or POST /api/v1/billing/checkout with { "plan": "pro" | "scale", "interval": "month" | "year" }. The response is { url }; open it to pay.
• Manage, change or cancel: from your dashboard, or POST /api/v1/billing/portal, which returns { url } for the Polar customer portal (payment method, invoices, cancellation). It works once you have paid for a plan.

Both endpoints need a signed-in browser session. API keys can't start a checkout.

Watching usage

• Your dashboard shows credits used and remaining, per-day totals and every request.
• GET /api/v1/usage?days=30 returns { plan, quota, totals, daily, events } (scope usage:read).
• anymd usage prints the same from the CLI; MCP has usage_summary.

Questions

Billing questions and Enterprise quotes: hello@digitop.ai. Refunds follow the refund policy.