AI AI Apps

Think of anything and it will find it in twenty yes/no questions - or let it hold the secret and you do the asking. Every question is one turn of one server-side conversation, and the whole established-fact ledger is restated on every turn so a long game cannot drift. When the app holds the answer it commits to it first: a code derived from the secret is shown before question one and checked against the reveal, so the answer cannot quietly become whatever fits your questions. When you hold the answer, the browser flags any question that repeats an earlier one, inverts it, or is already settled by it. A free lane needs no account at all - the fact ledger, the repeat detector, an information meter that shows the candidate space falling, a question-quality score, an axis coach, a two-player mode and two complete recorded games.

Put a question to a figure from history and get an answer in their voice — with every claim graded documented, attributed, inferred or beyond their world, so you can tell the record from the reconstruction. Only figures who died in or before 1905: no living or recently deceased people, no reciting their works, no invented present-day opinions. The grading is checked in your browser against how that figure's record actually survives, which is not the same for a Roman emperor's private notebook and a Lakota chief's words carried through an interpreter.

Build hosted mini-apps by chatting — describe an app and the assistant deploys it to {slug}.skillsafe.ai.

Generate photorealistic portraits of entirely fictional people, steered by broad attributes: age range, presentation, lighting, framing, lens, era, mood and treatment. Placeholder avatars, character references and mock-up faces with no real person behind them. Requests for the likeness of a real individual are refused, not attempted.

A battle simulator that adjudicates from a stat table, not from memory. Pick two contenders, set the terrain, light, preparation, victory condition and scale, and get a reasoned verdict: what each side brings, where the decisive advantage lies, how it most likely goes and what would flip it. Figures come from a curated stat table of 46 contenders that is passed into the request as fixed input, and every number in the verdict is audited back against it — so an invented bite force shows up on the page instead of passing as fact.

A surreal image generator you can actually steer. Choose the kind of strangeness — scale inversion, material substitution, physics violation, uncanny recombination, repetition, wrong place, inside out, absence, time out of joint, animacy swap — and the medium that carries it: oil painting, photographic realism, cut-paper collage, etched illustration and seven more. The brief is composed in your browser and shown in full before you spend anything. Invented scenes only; likenesses of real people and gore are refused before any credit is reserved.

Paste a rough AI prompt and get an optimized, self-contained rewrite built on proven prompting frameworks (RTF, RISEN, Chain of Thought, RODES and seven more), plus the gaps found, clarifying questions when details are missing, and a quality self-check. Free instant prompt lint runs in your browser. Derived from @sickn33/prompt-engineering.

Describe the capability you want an AI agent to have - or paste a SKILL.md you already use - and get a complete, well-formed agent skill: trigger-complete frontmatter, an imperative body under the 500-line budget, a bundled-resource plan, and an honest review that raises open questions instead of inventing answers. The SKILL.md linter runs free in your browser. Derived from @tldraw/skill-creator.

Chat with several AI models at the same time and compare their answers side by side. Every reply streams in with a receipt for what it actually cost - worst-case hold up front, charged at actual usage, the rest refunded. Pick up to four models per chat.

Paste a task spec and the deliverable produced against it. Two independent AI reviewers check it against your rubric with no shared context - both must pass before it ships. A NICE verdict means it ships; NAUGHTY returns a deduplicated fix plan ordered blockers-first and tagged by which reviewer caught what, plus a reviewer-agreement metric, a paste-ready fix prompt, and a re-check that diffs the revision criterion by criterion. Derived from @affaan-m/santa-method.

Turn a one-line objective plus your project facts into a dependency-ordered construction plan any coding agent can execute cold: self-contained context briefs, one-PR-sized steps, real verification commands, observable exit criteria, and parallel waves computed from the dependency graph. Paste an existing plan and the analyzer audits it free in your browser; forks only you can settle come back as open decisions, never guesses. Derived from @affaan-m/blueprint.

Paste your MCP server — TypeScript SDK, Python FastMCP or tool-definition JSON — and get a senior MCP engineer review: a readiness posture, the tool inventory, prioritized findings with corrected fragments, quick wins and focus areas. A free instant prescan flags missing descriptions and schemas, any-typed params, generic errors, unpaginated list tools, missing annotations, literal secrets, stdout logging on stdio and deprecated SSE before you run anything. Derived from @anthropics/mcp-builder.

Turn a sample API response into the response_semantics block and Adaptive Card templates that make Microsoft 365 Copilot render your plugin data as cards. The free half is a real engine: paste the JSON and it measures every field with type, sample and presence count, generates a grounded skeleton bundle, validates any bundle against the card rules AND against the sample — a binding that resolves on no item is named — and previews the exact card each item would produce. The AI pass shapes the template around your intent, and the same validator re-checks every binding it used, in public. A derived work of @github/mcp-create-adaptive-cards (github/awesome-copilot).

Paste a draft email, message, listing, or report. Get it back without AI tells: no filler openers, no 'it's not X, it's Y', no punchy fragments, active voice, complete sentences. Keeps every fact; lists what it changed.

Ask an AI image model to mark up a photo and it returns a subtly re-rendered image rather than your original. This isolates just the red annotations as a transparent PNG - with a sensitivity slider - and can composite them back onto your untouched original, so you get clean marks over pristine pixels. Runs fully client-side; neither image is uploaded. Imported Claude artifact. Source: https://github.com/simonw/tools (red-extractor.html, Apache-2.0).

Paste a vision model's segmentation output - objects with box_2d bounds and base64 masks - and see the boxes and masks drawn over the image, with a switchable co-ordinate origin for when a mask lands mirrored or rotated. Ships an embedded example so it demos without an upload. Runs fully client-side. Imported Claude artifact. Source: https://github.com/simonw/tools (mask-visualizer.html, Apache-2.0).