A: 1. Credit costs. We moved to a single credits meter, which is why the old "premium + assistant" split reads as confusing. Current costs, charged once per action regardless of how many model calls happen under the hood (a single generation fans out to about five internal calls and still costs 1):

Generator prompt: 1 credit
Optimizer pass: 1 credit
Chat refinement / refine an existing prompt: 0.1 credits
Prompt Tester message: 1 credit per message on the current default model roster

Tester is priced per model, so premium models cost more when we add them, and the cost is shown next to the model before you send. Monthly allotments: Free 5 credits, Tier 1 1,000, Tier 2 3,000, Tier 3 12,000,
Tier 4 30,000. Your exact remaining balance and reset date are in the sidebar under Usage. At 0.1 per refinement, refinement-heavy work is very cheap.

2. Data handling. Straight answers, all of this is written into our Privacy Policy (promptbuilder.cc/privacy-policy, Sections 3 and 4):

Encryption: TLS in transit, AES-256 at rest.

Storage location: United States. Vercel (US East) for hosting, Neon Postgres (AWS us-east-2) for account data, saved prompts, and history, MongoDB Atlas (US) for operational data, AWS SES (US West) for transactional email.

DPA: Yes. A signed GDPR DPA is available on request, email [email protected]. The full subprocessor list is published in Section 3 of the Policy, and we update it with the "last updated" date whenever it materially changes.

Model provider retention: We use Anthropic, OpenAI, Google, DeepSeek, and xAI through their commercial APIs, with some routed via OpenRouter. Under those API terms, content submitted is not used to train their models. Providers may retain API content for a limited window, typically up to 30 days, for abuse monitoring, then delete it per their policies. We link each provider's data-usage policy directly in the Policy so you can verify rather than take our word.

Us: We do not train our own models on your data, we do not use it for ad profiling, and we do not sell it. EEA/UK/Swiss transfers are covered by Standard Contractual Clauses, and you have full access, rectification, erasure, portability, and objection rights.

3. Roadmap. Already shipped since launch: Library folders (create, rename, delete, bulk move, deleting a folder never deletes prompts) and team workspaces with roles and invites. So the folders ask is done.

Next up: interactive Optimizer with guided improvements and comparisons, prompt sharing via links, an image prompt library, image generation flows, and API access for generation, optimization, and library sync. We ship continuously and post every release to the changelog at docs.promptbuilder.cc/changelog.

4. The "Unlimited" tier. Fair question, and you're right that the wording deserves clarity. It means 12,000 credits per billing cycle, which resets monthly, forever. That's 12,000 generations or optimizations a month, or 120,000 refinements. If you did hit it, generation pauses until the cycle resets rather than silently degrading, and our Terms reserve the right to apply rate limits past that point to protect the service from abuse. In practice essentially nobody reaches it through normal use, and if you have a legitimate workload that does, email [email protected] and we'll work it out with you rather than leave you stuck.