PromptOps
Engineer prompts like software.
A native iOS studio for building, testing, and shipping AI prompts. Version every change, score it with real evaluators, and run it against any model you like — all on your device, with your own API keys.
The problem
Your best prompts live in a Notes app, a Slack thread, and a forgotten chat window. When you tweak one, the old version is gone. When you switch models, you start from scratch. When someone asks whether the change actually made it better — you eyeball it.
Prompts are infrastructure now. They deserve the tools that infrastructure gets.
The approach
PromptOps treats prompts like code. Every edit creates a version. Every version can be scored against a dataset. Every supported model is one tap away, and every result is a number you can compare.
It runs natively on iOS 26, stores everything on device, and uses your own API keys — so the people who built it have no way to see what you are working on.
A real prompt workflow, in your pocket
Everything a serious prompt practice needs — authoring, evaluation, versioning, and comparison — without a browser tab in sight.
Multi-provider, one workflow
OpenAI, Anthropic, Google Gemini, xAI, Ollama, and any OpenAI-compatible endpoint. Swap the model without rewriting the prompt, and diff the outputs side by side.
Version every change
Draft, publish, deprecate, roll back. A full history per prompt with instant restore. You never overwrite a prompt that was working.
Score it, don't guess
16 built-in evaluators — json_valid, schema_valid, regex_match, required_fields, similarity, latency, cost, tool-call checks, and model-graded rubrics with weighted dimensions.
Variables with guardrails
{{variable}} placeholders with types, defaults, required flags, min/max length, regex patterns, and enum validation. Bad inputs fail before they reach a model.
47 production templates
Writing, coding, research, data, agents, RAG, vision, safety, and design. Clone one and it arrives with a bespoke evaluation suite already attached — 51 suites ship in the box.
Chain prompts into Skills
Bundle prompts into ordered chains where each step feeds the next, or fan them out as a parallel batch. Seven ready-made Skill blueprints install on demand.
See what actually changed
Score trends per suite, KPI cards, regression alerts, and failure clustering that groups losing cases and suggests a fix. Charts, not vibes.
Inspect the exact request
The Prompt Inspector renders every message turn, tool, setting, and cache breakpoint you are about to send, with per-block token estimates. No more guessing what the model received.
Built for iOS 26
SwiftUI throughout, Liquid Glass chrome, instant cold start, Dark Mode, full Dynamic Type and VoiceOver support. iPhone and iPad. Native, not a wrapper.
Four steps from blank prompt to shipped
The same loop you already use for code, applied to the text that drives your models.
Design
Start from a template or a blank prompt. Add variables and validation rules. Pick a model. The built-in checkup flags missing system prompts, unused variables, and JSON output with no schema.
Run
Execute it with streaming, thinking, caching, and structured outputs where the model supports them. Inspect the composed request. Save a version as you go.
Evaluate
Build a test dataset, attach evaluators, run the suite. Get a pass rate per version, a per-dimension rubric breakdown, and clusters of what failed and why.
Ship
Publish the version that won. Validate it against a holdout set so you know you did not overfit. Export to JSON or sync to a repo you own.
The core studio is free
Authoring, execution, versioning, templates, and the full evaluation engine are included. Pro adds the team-scale workflows.
The full studio
Everything you need to write, run, and score prompts properly.
- Unlimited prompts, versions, and executions
- All 47 production templates and 51 eval suites
- All 16 evaluators, datasets, and suite runs
- OpenAI, Anthropic, Gemini, xAI, and Ollama
- Prompt Inspector, checkup linter, and token counting
- AI authoring tools — Prompt Improver and Test-Case Generator
- Insights charts and failure clustering
- Full JSON export
Scale the loop
An in-app subscription for the workflows that span many prompts and many runs.
- Skills — ordered chains and parallel batches
- Compare — two models or two versions, diffed side by side
- Regression Check — replay every prior run against a new version
- Sync — push and pull your library to GitHub or iCloud
- Custom providers — any OpenAI-compatible API
- Batch evaluation — suites run async at reduced cost on supported models
Current pricing, trial terms, and Family Sharing details are shown on the App Store listing.
Built for
- Developers wiring LLMs into production apps
- Prompt engineers iterating on prompts that have to work every time
- Product teams pressure-testing an AI feature before committing to it
- Researchers benchmarking models against their own data
- Anyone tired of pasting prompts between tabs
Why people pick it
- Bring your own key. No markup on inference. You pay your provider directly, at their rates.
- No lock-in. Export the whole database to JSON whenever you want, or sync to a repo you control.
- No backend means no breach. There is nothing to compromise on a server, because there is no server.
- Native, not a wrapper. Built for iOS from the first line, not a web app in a shell.
- Current-generation model support. Adaptive thinking and effort, prompt caching, structured outputs, and few-shot examples, wired properly per provider.
Privacy that isn't a marketing line
PromptOps has no backend, no account system, and no analytics SDK. When you run a prompt, your device makes the API call directly to the provider you configured. We did not just promise not to look — we built it so there is nothing to look at.
On device
Prompts, versions, datasets, and run history are stored locally in SwiftData.
In the Keychain
API keys are held in the iOS Keychain and sent only to the provider they belong to.
No telemetry
No analytics, no crash-reporting SDK, no tracking identifiers, no ads.
Yours to take
Export everything to JSON, or sync to your own GitHub repo or iCloud Drive.
Before you download
Do I need an account?
No. There is no sign-up, no login, and no server to register with. Install the app and start writing.
Do I need my own API keys?
Yes. PromptOps is bring-your-own-key: you add a key from OpenAI, Anthropic, Google, xAI, or any OpenAI-compatible provider, and the app calls that provider directly from your device. You are billed by the provider at their rates, with nothing added on top. If you run Ollama on your network, you can use local models with no key at all.
Which models does it support?
OpenAI, Anthropic, Google Gemini, xAI, Ollama, and any OpenAI-compatible endpoint — Groq, Together, Fireworks, Azure, or your own. Where a model supports streaming, extended thinking with an effort setting, prompt caching, structured outputs, or tool use, PromptOps wires those through rather than flattening everything to plain text.
What devices does it run on?
iPhone and iPad, running iOS 26 or later. It is a universal app — one download covers both.
Where does my data live?
On your device. Prompts, versions, datasets, and execution history are stored locally; API keys live in the iOS Keychain. Nothing is sent anywhere except the API request you explicitly run, which goes straight to the provider you chose.
Can I get my data out again?
Any time. Export the full library to JSON from Settings. With Pro you can also push and pull to a GitHub repository you own, or sync through your own iCloud Drive.
What exactly does an evaluation do?
You attach evaluators to a suite and point it at a test dataset. The suite runs your prompt over every case and scores each one — deterministic checks like JSON validity, schema conformance, regex matches, required fields, and length bounds, plus similarity scoring and model-graded rubrics with weighted dimensions. You get a pass rate, a per-case breakdown, and a comparison against previous versions.
How do I get help?
Read the support page or email abriggs@riteupai.com. Include your iOS version and the PromptOps version from Settings → About and you will get a faster answer.
See it
A prompt IDE that fits in your hand.
How it works
Four steps from blank prompt to production-ready output.
Design
Start from a template or a blank prompt. Add variables. Pick a model.
Test
Run it. Tweak it. Save versions as you go. Compare outputs across providers in seconds.
Evaluate
Build a test dataset. Attach evaluators. Run the suite. See pass rates per version.
Ship
Publish the version that works. Sync to GitHub. Use it in your app.
Built for
- Developers integrating LLMs into production apps
- Prompt engineers iterating on prompts that need to actually work
- Product teams experimenting with AI features before committing
- Researchers benchmarking models on their own data
- Anyone tired of pasting prompts between tabs
Why people pick PromptOps
- Bring your own key. No markup, no margin on top of your provider bill. Pay OpenAI or Anthropic directly, at their rates.
- No lock-in. Export the full database to JSON anytime. Sync to a GitHub repo you own. Take everything with you.
- No backend means no breach. There is nothing to compromise on a server because there is no server.
- Native, not a wrapper. Built specifically for iPhone and iPad, not a port of a web app.
Simple pricing
Free does real work on its own. Pro unlocks the workflow features.
Free
Everything you need to author, run, and evaluate prompts on your own.
- Prompt authoring with full versioning
- 47 seeded templates across 13 categories
- Single-prompt execution against any configured provider
- 12 rule-based + 2 model-graded evaluators
- Eval suites, datasets, and runs
- Full JSON export of your database
- No usage cap, no ads, no trial expiry
Pro
Or $39.99 / year (~33% off). 7-day free trial on monthly. Family-shareable.
- Skills — chained or bundled prompt flows
- Side-by-side compare across versions or providers
- Regression check — replay prior runs against new versions
- Sync to GitHub or iCloud, including Skills
- Custom OpenAI-compatible providers beyond the five built-ins
- All Free features included
Pro is an Apple In-App Purchase managed by your Apple ID. Cancel anytime in Settings. Provider API costs are billed directly to you by OpenAI, Anthropic, etc. — PromptOps adds no markup.
Privacy that's not a marketing line
PromptOps has no backend. No account system. No analytics SDK. When you execute a prompt, your iPhone or iPad makes the API call directly to the provider you chose. We never built the means to see your data.
Ready to engineer your prompts?
Free to download. Optional Pro at $4.99/mo or $39.99/yr.