← Back to CreditFX

Changelog

What's changed, added, or corrected on CreditFX — newest first. Pricing corrections and new platforms happen regularly; this is where they're logged.

09.03.26

A split-flap animation, and workload presets (Visla, as a proof of concept)

  • The Usage Planner's recommended cost now does a real departure-board style flip when it changes — switching modes, adjusting quantity or length, or picking a preset all trigger it. Used deliberately in one place only, not applied broadly across the site, since the motion is meant to mark a moment that actually changed, not decorate every number.
  • Added "typical workload" preset buttons above the Usage Planner's quantity inputs — tap one to instantly fill in a realistic starting point instead of guessing your own numbers from scratch. Presets are mode-aware: switching between a platform's different creation modes shows a different, relevant set.
  • Currently live for Visla only, with placeholder numbers meant to prove the mechanism feels right before writing real, considered presets platform by platform — same verification discipline as the rest of the dataset, not a copy-paste pass across all 8 Usage Planner platforms.
08.31.26

Mobile ticker fixes — actually scrolling now

  • Fixed the scrolling rate ticker not visibly moving on mobile — it had no explicit white-space rule, so on a narrow screen the text wrapped onto multiple lines instead of staying on one scrolling row, defeating the whole effect.
  • Also fixed the ticker's speed being tied to a fixed duration, which meant it silently got faster every time a platform was added — the same time window then had to cover more distance. Speed is now measured from the actual content width and holds a constant pace regardless of how large the dataset grows.
  • If the ticker still looks static on a specific device after this, worth checking that device's Reduce Motion accessibility setting — CreditFX intentionally stops the animation when that's enabled, which is correct behavior, not a bug.
08.30.26

5 new platforms, and ElevenLabs joins the Usage Planner

  • Added Cartesia, Murf, and Gradium (Voice & Audio), Azure AI Speech (cloud API, pay-as-you-go), and CheaperInference (a discounted model-router marketplace, same category as OpenRouter) — all verified from public pricing pages, no screenshots needed.
  • Caught a real, very recent price change along the way: OpenAI cut GPT-5.6 Sol's output rate from $30 to $20 per million tokens on August 22 — confirmed directly against OpenAI's own developer docs. The other two tiers in the family were already correct.
  • ElevenLabs joins the Usage Planner (Limited confidence) — its overage rate is confirmed directly from ElevenLabs' own API pricing page, resolving several conflicting third-party numbers found during research.
  • Fixed the Usage Planner's input wording, which had "videos" hardcoded for every platform — accurate for the video-generation platforms already live, but wrong for ElevenLabs (an audio tool). Now configurable per platform, so this won't need a one-off fix again as more non-video platforms get added.
08.29.26

A feedback form and a WhatsApp community

  • Added a link to a short feedback form, in the footer (every page) and on the Support page.
  • Added a WhatsApp community — join to hear about new platforms and features before they're released. Linked directly from a new FAQ entry on the Support page.
08.25.26

Usage Planner is easier to find and faster to open

  • Made the Usage Planner much easier to discover across CreditFX. A new "NEW · PLANNER" badge now appears anywhere the planner is available, using a darker pink treatment for better readability.
  • The new badge is also a shortcut — click it from the platform list or an individual platform page and CreditFX takes you directly to that platform's Usage Planner with it already open.
  • Added a callout inside expanded platform rows to make it clear when a Usage Planner is available, with a direct "Plan usage" action instead of requiring users to find the feature themselves.
  • Added a one-time tooltip introducing the Usage Planner to new visitors. Once dismissed, it stays dismissed and won't keep interrupting you as you browse other platforms.
08.24.26

Usage Planner grows to 7 platforms, zero screenshots needed

  • Added D-ID Studio, Hedra, and Colossyan to the Usage Planner — verified entirely from public pricing and help-center pages, no account screenshots required.
  • Caught a real, unrelated find along the way: Colossyan's pricing had genuinely changed since it was last checked (Starter went from $19/mo to free; the paid tier from a $70 'Business' plan to a $59 'Professional' plan). Not something we were looking for — it just turned up during verification.
  • Two new honest limitations, each clearly labeled rather than glossed over: Hedra's credit top-up packs are confirmed to exist but their exact rate isn't published, so over-budget workloads show as unsupported even though buying more is actually possible; Colossyan's older and newer avatar models draw from two completely separate credit pools, so only the current model is modeled for now.
  • 7 platforms now live: Visla, Runway, Synthesia, HeyGen, D-ID Studio, Hedra, Colossyan.
08.23.26

Usage Planner — plan real workloads, not just single calculations

  • New feature: on supported platform pages, "Plan your usage" now takes plain-language input (how many videos, how long each) and checks every real pricing tier — not just the one you'd think to pick — to find the actual cheapest option for that workload, including top-up costs where they apply.
  • Every recommendation carries a confidence label. "Exact" means the consumption rate is directly confirmed; "Limited" means only some of a platform's features are modeled, with the specific gap stated in plain language rather than glossed over.
  • Live for four platforms so far, each independently verified against real sources before being turned on: Visla, Runway, Synthesia, and HeyGen. Two caught real surprises along the way — a cheaper annual-billing tier beating every monthly option on Visla, and Synthesia's complete lack of any overage-purchase option (exceed your credits and you upgrade or wait, no exceptions).
  • One data error caught and fixed in the process: a platform had a 'standalone image generation' mode that doesn't actually exist as a real product feature — removed rather than left in under a false confidence label.
  • Also fixed this round, from an external code review: a Leonardo AI entry that could rank as literally $0/image and distort its category's leaderboard, a Privacy Policy paragraph that no longer matched what the code actually does, and restored keyboard focus outlines that had been globally disabled.
08.22.26

Tried a chart view, removed it, kept the fix it uncovered

  • Added an experimental scatter-chart view (rate vs. monthly cost) for the Video Generation category, using a log scale since AI pricing tends to cluster a lot of affordable options next to one or two expensive outliers — a plain scale crushed that cluster unreadable.
  • While extending it, found and fixed a real gap: platforms priced per seat/month had no comparability logic at all, so they were invisible to every leaderboard, not just the chart. That fix is staying — it's why Editing & Repurposing and AI Writing now have real leaderboards they didn't have before.
  • Even after the fix, the chart only ever qualified for 3 of 11 categories (needed 4+ platforms sharing both a real monthly cost and a comparable unit). A feature that sometimes shows up and sometimes doesn't creates more confusion than it's worth, so it's been removed — consistency over the chart itself.
08.21.26

Individual platform pages, dataset expansion, and a routing bug fix

  • Fixed a real bug from yesterday's prerendering work: a catch-all routing rule was quietly serving the generic homepage shell instead of the real page content on About, Support, and the other static pages. Fixed and verified directly against the live raw HTML.
  • Every platform now has its own real page and URL (e.g. creditfx.mogabr.com/platform/runway) — 107 pages, each with its own title, description, and structured data for search engines and AI answer tools.
  • Added 17 new platforms: xAI, Mistral AI, Cohere, AI21 Labs, Perplexity, Voyage AI, OpenRouter, Together AI, SambaNova Cloud, DeepInfra, Cloudflare Workers AI, Google Vertex AI, Azure AI Foundry, Replicate, fal, Modal, and Hugging Face Inference Endpoints.
  • Declined to add Banana (confirmed shut down in 2024), NVIDIA NIM (no public per-token pricing), and IBM watsonx.ai / Oracle OCI (couldn't verify current self-serve pricing with confidence).
  • Updated Kimi and Qwen to their actual current flagship models (K3 and Qwen3.8-Max) — both had moved past what was previously listed, and Qwen is now the cheaper of the two, a reversal from before.
  • Named the specific model or resolution behind the rate for 7 platforms that were previously ambiguous: Pika, Hailuo, Adobe Firefly, Dreamina, DeepSeek, GLM, and Cloudflare Workers AI.
  • Added full tier breakdowns (previously only one tier each) for Pika, Luma AI, Hailuo, Midjourney, and ElevenLabs.
  • Added a cross-reference for platforms whose editing/repurposing capabilities were previously undiscoverable from that category tab (Visla, Vyond, Powtoon, InVideo) — points to the real entry rather than duplicating it with an invented rate.
  • Added Kimi and Qwen to the automated price-watch list, so a repeat of the above model-drift issue gets caught automatically going forward.
  • Increased font size on platform descriptions and reviews for easier reading.
08.20.26

Search & AI-answer visibility

  • About, How It Works, Changelog, Privacy, Terms, and Support now deliver real, readable content immediately — not just after the page's JavaScript finishes loading. This directly affects how search engines and AI answer tools discover and reference CreditFX.
  • Added structured data (machine-readable facts search engines and AI tools can read directly) to the homepage, About, and Support pages.
  • Added a sitemap and explicit crawl permissions so every page can actually be found.
  • Increased the font size on platform descriptions and reviews for easier reading.
  • Reduced overuse of the brand color across the board so rates and the top pick in each category stand out more clearly.
08.19.26

Launch

  • Launched CreditFX — 90+ platforms across 11 categories: Video Generation, Editing & Repurposing, Image Generation, Voice & Audio, Music Generation, Coding & Agents, LLM APIs, GPU Cloud, AI Memory, AI Writing, and Stock + AI.
  • Category leaderboards that only rank platforms sharing a genuinely comparable pricing unit — mixed billing models are shown separately, never blended into one misleading ranking.
  • Multi-tier and multi-mode calculator support for platforms with several pricing plans or several ways to create (Visla, Runway, Synthesia, HeyGen, D-ID Studio, Hedra, and Colossyan so far — more being added over time).
  • Save calculations across multiple platforms and export them together as one branded PDF, with a running total.
  • Full dataset available as a free CSV download, plus an opt-in to hear about pricing changes.
  • Search across all 90+ platforms by name.

Notice a rate that's changed or a platform that's missing? Head to the Support page — corrections and suggestions both feed directly into future entries here.