Loading Jonarix?
Dashboard

Home

Your prompts and registry activity. Use / in the header to search.

My prompts
0
0 published ? 0 draft
Total Runs
0
across all prompts
Plan
Free
Creators to watch

Post prompts with screenshots or short demos, earn stars, and grow your presence in the community catalog.

Recent Prompts

Name Status Model Updated

No prompts yet

Create your first prompt to start versioning your AI work.

Quick Actions

Activity

Welcome to the Jonarix registry. Open the community catalog or publish your first prompt when you are ready.
Just now

Your prompts

Manage, version, and publish your AI prompts.

Prompt Catalog / quality Version Models Runs Stars Updated

No prompts yet

Your versioned prompts live here. Create one and start building your prompt library.

Select a prompt from the catalog ?

Starred prompts

Your starred picks from the community catalog ? quick access to prompts you trust.

No stars yet

Browse the community and star prompts you want to save or come back to.

Developer directory

Authors who ship in the public catalog. Search by @handle, name, or focus area ? then jump to their prompts.

Tip: press / in the header to focus global search ? your query can open this view with the same text.

Developer profile

Public profile

Dashboards

Visual prototype for the in-app dashboard builder ? layout reference only; controls are not wired yet.

App dashboard UI prototype
Dashboard shell reference
Dashboard layout variant reference
Dashboard management prototype reference

Public profile

Your developer home on Jonarix: a GitHub-style public page with sections, rich links, availability, and a full README preview. Your @handle surfaces in the catalog and CLI once you publish.

Build your registry page like a GitHub profile: rich fields, social graph, and a full-width README preview.

Public presence

Headline and bio appear at the top of jonarix.com/@handle. Display name is managed under Account settings.

Use an https:// image URL, or a compact data:image/? data URI. Preview confirms the image loads; remote hosts must allow hotlinking.

Appearance & theme

Colours apply to your public handle page and this preview: banner gradient, accents on chips and links, and avatar glow.

#A855F7

Buttons, chips, and highlights.

#2563EB
#16A34A

Work & availability

Company and location show as metadata. Use availability for consulting, collabs, or office hours.

Social & links

Add every surface where people can verify you or reach out. URLs are normalised to https.

Showcase your work

Pin up to 12 published prompts on your public page. Visitors see cards with name, badge, and models ? ideal for highlighting your best packs.

Loading your prompts?

README

Long-form story: projects, principles, and how to use your prompts. Supports headings (##), bullets, **bold**, *italic*. Scripts are stripped in preview.

0 / 20,000 characters

Visibility & publish

Save stores drafts on your workspace. Publish profile marks this page as live so people who open your handle from the community catalog see the latest version.

Not published yet ? click Publish when you are ready for visitors.

API Key

Manage runtime API keys for CLI, studio integrations, prompt runs, and agent scans.

How runtime keys work (architecture summary)
  • Generation: Create runtime keys directly from this page.
  • Control: You can create multiple keys and revoke any key instantly.
  • Usage: Copy one active key to your CLI or studio environment.
  • Execution path: your IDE/CLI -> runtime auth -> model endpoint.
  • Operations: Rotate keys regularly and revoke unused keys immediately.

Environment Profiles

Select a profile to prefill key metadata and recommended access policy.

No profile selected. Choose one profile, then generate a key to apply its policy.
# Runtime identity (technical profile)
service_name   = "jonarix"
cli_binary     = "jonarix"
auth_header    = "Authorization: Bearer <runtime_key>"
env_var        = "JONARIX_API_KEY"
default_mode   = "production"

HTTP API ? connect agents & IDEs

Use the same runtime key you generate below (jnx_live_?) over HTTPS so tools, scripts, and coding agents (for example Cursor with a project rule) can talk to Jonarix without your Firebase login. This is how external models list your workspace prompts and load the full prompt body, including the latest text from version history when needed.

Base URL for this site: https://jonarix.com/api

  • GET /health ? No auth. Returns ok when the API is reachable (via Netlify /api/* proxy or Firebase Hosting).
  • POST /runtime/verify ? Optional key check. Send JSON {"key":"<jnx_live_?>"} or header Authorization: Bearer <key>. Response includes keyId and status.
  • GET /runtime/prompts ? Requires Authorization: Bearer <runtime_key>. Lists prompts you own (id, name, description, bodySnippet, updatedAt, ?).
  • GET /runtime/prompts/<promptId> ? Same Bearer header. Returns the full prompt object including body, description, planMarkdown, codeSnippet, and related fields.

Set the key only in environment variables (for example JONARIX_API_KEY). Never paste runtime keys into public repos, client bundles, or shared prompts. Rotate a key from this page if it is ever exposed.

# List prompts (set JONARIX_API_BASE to your API root, e.g. https://jonarix.com/api)
curl -sS -H "Authorization: Bearer $JONARIX_API_KEY" \
  "${JONARIX_API_BASE}/runtime/prompts"

# PowerShell example:
# Invoke-RestMethod -Uri ($env:JONARIX_API_BASE + '/runtime/prompts') -Headers @{ Authorization = "Bearer $($env:JONARIX_API_KEY)" }
# Example: fetch one prompt by Firestore document id
curl -sS -H "Authorization: Bearer $JONARIX_API_KEY" \
  "https://jonarix.com/api/runtime/prompts/<promptId>"

For agents: tell the model the base URL, that every protected route uses Bearer auth, and the workflow ?list ? pick id ? GET detail ? use prompt.body?.

Generate & Connect Runtime Key
Jonarix generates keys and lets you create, rotate, and revoke them anytime
No keys

Quota & Reliability

API Key Usage Analytics

Focused analytics on trend, source quality, and recent key usage context.

Usage trend (last 7 days)
Usage by source
Recent usage (where key was used)

External Runtime Ingestion Spec

Use this payload schema from external studios/agents so usage appears in the same analytics dashboard.

# POST usage event (studio/agent integration)
{
  "keyId": "<runtime_key_id>",
  "source": "external.studio.run",
  "where": "vscode/linux",
  "platform": "Win32",
  "online": "online",
  "status": "success",
  "errorCode": "",
  "latencyMs": 420,
  "tokensIn": 1200,
  "tokensOut": 340,
  "createdAtMs": 1730000000000
}

API Key FAQ

Runtime keys are stored in your Jonarix account scope and shown in this dashboard list. Only keys with active status should be used in local tools.

  1. Generate a new key.
  2. Update CLI/studio to use the new key.
  3. Run a quick test command.
  4. Revoke the previous key.

Yes. Create separate keys for different environments (local development, CI, automation) so access can be rotated or revoked independently.

  • Key status is active in this dashboard
  • The same key is set in both editor and terminal sessions
  • No stale environment variable is overriding your expected key
  • You are signed into the intended Jonarix account/workspace

Using key with CLI and studio (floating guide)

Use one active key per working environment. This keeps install, run, and agent scan behavior consistent and avoids cross-environment permission drift.

  • Generate in Jonarix: Keys are created by system logic and bound to your signed-in account.
  • Distribute intentionally: Copy only active keys to trusted local CLI/studio environments.
  • Rotate safely: Create a replacement key, verify it, then revoke old keys.

Run this exact sequence:

# 0) Click "Generate key" above (saved to your account)
npm install -g jonarix
jonarix keys set <generated_key>
jonarix keys status
jonarix install <handle/prompt>
jonarix run <handle/prompt>
jonarix agent scan --prompt <handle/prompt> --path .

# Optional override for current shell session only
$env:JONARIX_API_KEY="<generated_key>"

Production verification checklist:

# Must return the expected account + active key
jonarix whoami
jonarix keys status

# Must install and execute successfully
jonarix install <handle/prompt>
jonarix run <handle/prompt> --var smoke_test=1

# Must complete agent workflow and return output
jonarix agent scan --prompt <handle/prompt> --path . --output json

Operational rule: one selected active key per environment, explicit rotation, immediate revocation for unused keys. If studio install works but agent scan fails, verify user/session alignment, key freshness, and command context first.

Workspace

Account & workspace

Identity, public handle, billing limits, security, and optional catalog verification.

Profile, workspace URL, plan limits, and session options.

Profile

Name and email shown in the app and catalog.

?

?

Uses your public profile photo when set; otherwise initials.

Shown on your public profile and in the catalog.

Read-only. Change it with your sign-in provider, then sign in again.

Workspace

Editable handle and public URL (must be unique).

Use 3-32 chars: letters, numbers, hyphens, underscores.

Public URL jonarix.com/?

Registry & discovery

Discovery and your public developer page.

Public developer page

Edit headline, README, links, and preview before publishing.

Your sign-in email is not shown publicly unless you add it to your README.

Appearance

Choose your preferred color theme.

Currently: Dark

Security

Password reset and two-factor authentication (TOTP).

Password

If you use email and password to sign in, we can send a reset link to your account email.

Two-factor authentication (2FA)

Google Authenticator, Authy, Microsoft Authenticator, or any TOTP app.

Status Off

Authenticator MFA must be enabled for Jonarix sign-in. If setup returns ?operation not allowed?, an administrator needs to turn on TOTP in the cloud auth console.

Plan & usage

Limits and upgrade.

Current plan
Free
Prompt usage

Free plan: 10 prompts

0 / 10 prompts 0%

Session & account termination

Sign out here or permanently delete your data.

Sign out

Ends your session on this browser. Other devices stay signed in until you revoke them.

Delete account

Permanently removes your account, prompts, and registry data. This cannot be undone.

LIMITED OFFER 100% OFF — Every feature is completely free during our launch. Expires in 21 days — No credit card required.

Plans & billing

During our launch promo, every feature — unlimited prompts, private repos, analytics, the full academy — is 100% free for all users.

Free

$0forever

Includes

  • 10 prompts in your workspace
  • Community catalog & stars
  • CLI & local BYOK keys
  • Public profile & handle

Team

$49/ month ? 5 seats

For small orgs

  • Shared org repos & admin tools
  • Higher pooled run volume
  • Team token & onboarding

Enterprise

Custom SSO ? SLA

Security & scale

  • SSO / SAML and compliance exports
  • Volume pricing & review
  • Optional self-hosted scoping

Full marketing comparison also lives on jonarix.com/pricing.

Billing history

Payments, invoices, and receipts for your workspace. Coffee support and plan upgrades are saved automatically after successful payment.

No payments recorded yet. After you subscribe, each charge appears here with invoice and receipt links when available.

Coding Academy

48 interactive lessons — solo coding or vibe coding with AI.

Full access

Pick your track and start building.

48 lessons across two paths — type every line yourself, or build fast with AI assistance. Both tracks are completely free.

Voice-narrated lessons
Live code editors
AI-assisted track
Completion certificate
100% free — always

Help & Documentation

Everything you need to get the most out of Jonarix ? from first prompt to production.

Getting Started

Jonarix is a platform for versioning, sharing, and installing AI prompts ? the same way npm handles JavaScript packages and Git handles code.

Every prompt you create gets a semantic version number (like 1.0.0), a stability badge, a variable schema, and a community test record. You can publish prompts so others can install and run them with a single CLI command.

Think of it as your team's shared prompt library ? searchable, versioned, and always traceable back to who wrote it and what results it produces.

Click + New Prompt (header or sidebar). The creator opens in the main panel ? a four-step flow:

  • Step 1 ? Basics: Kebab-case name, a primary category (kind of task), optional topic tags (Travel, SaaS, Entertainment?), and visibility.
  • Step 2 ? Details: Description, optional text examples, and a required demo ? one screenshot is enough. You may add more images if you want (optional, max 5), or use one short video (up to 5 seconds) instead of images.
  • Step 3 ? Prompt: Your actual instructions for the model. Use {{variable}} for placeholders ? they're auto-detected. Pick model compatibility.
  • Step 4 ? Publish: Initial version, then save as draft or publish.

Once published, anyone can install and run it using jonarix run your-handle/prompt-name.

This is the full quality workflow used by top-performing prompts in the catalog. Follow it from start to finish to publish something people can actually run, trust, and star.

Before you click New Prompt (2-minute prep)

  • Write the exact user problem in one sentence (example: "Turn messy meeting notes into action items by owner and due date").
  • Collect 2-3 real input samples and define what "good output" looks like.
  • Decide your target models first (the model chips in Step 3 should reflect real tests, not guesses).

Step 1 ? Positioning (discoverability + routing)

  • Prompt name (newPromptName): use kebab-case, action-first names such as summarize-quarterly-report. Avoid generic names like assistant. The input is cleaned by sanitizePromptName().
  • Primary category (newPromptCategory): pick the core task type users will filter by in Community Catalog.
  • Visibility (newPromptVisibility): choose audience mode for published prompts (public/private).
  • Publish mode (Step 4 radios): choose Save as draft or Publish now. Final visibility is resolved in submitNewPrompt().
  • Topics & verticals (newPromptTopicsGrid + newPromptTopicOtherDesc): add industry tags users search for. "Other" field visibility is controlled by syncPromptTopicOtherVisibility().

Step 2 ? Context (trust + conversion)

  • Description (newPromptDesc): include who it is for, expected input shape, and exact output format. This text decides whether users click your prompt.
  • Example input/output (newPromptExampleIn, newPromptExampleOut): show one realistic pair. This reduces misuse and support questions.
  • Demo preview (required): upload at least one image or one video. Files are managed through demo handlers and previewed before publish.
  • Negative examples (newPromptNegativeExamples): document bad outputs to prevent common failure modes.
  • Output shape (newPromptOutputSchema): define structure (JSON schema or strict object template) when consistency matters.
  • License + attribution (newPromptLicense, newPromptLicenseNote): clarify reuse rights so teams can adopt your prompt safely.

Step 3 ? Blueprint and body (execution quality)

  • Blueprint fields: role, inputs schema, assumptions, constraints, output format, fallback rules, quality checklist, and optional few-shot examples.
  • Insert blueprint: the button runs insertPromptBlueprintFromFields() and appends a structured template into your main prompt body.
  • Main prompt body (newPromptBody): write deterministic instructions with clear priorities ("If conflict, follow section X first").
  • Variables: use {{variable}} placeholders. They are auto-detected via extractVariables() into varsList.
  • Model compatibility: select only models you actually validated. Wrong model tags lower user trust fast.
  • Advanced docs (setup/code/plan): optional but strongly recommended for technical prompts and reproducibility.

Step 4 ? Review & publish (release discipline)

  • Version (newPromptVersion): start with 1.0.0 for first public release.
  • Changelog (newPromptVersionChangelog): explain what changed and why.
  • Codegen acknowledgment: certain categories require confirmation; visibility toggling is handled by syncCodegenAckHintVisibility().
  • Create prompt: submit runs submitNewPrompt(), writes the prompt document, then writes version 1.0.0 in prompts/{id}/versions.
  • Post-publish refresh: if public, the flow reloads prompts and community lists so your item appears in your tables and catalog views.

What each important function does

  • sanitizePromptName(input): enforces clean handle-style naming.
  • syncPromptTopicOtherVisibility(): toggles custom topic input visibility.
  • insertPromptBlueprintFromFields(): converts blueprint form fields into a structured body block.
  • extractVariables(): parses {{variables}} and updates the variable list.
  • syncCodegenAckHintVisibility(): shows required policy acknowledgment for sensitive categories.
  • submitNewPrompt(): validates fields, builds payload, uploads metadata, writes prompt + first version, and refreshes UI datasets.
  • loadMyPrompts(): repopulates your "Your prompts" table after creation.
  • loadCommunityCatalog() + renderDiscover(): refreshes public catalog cards when prompt is published.

High-quality prompt checklist (copy this before publish)

  • One clear use-case sentence in description.
  • At least one realistic demo media file.
  • At least one real input/output pair.
  • Explicit constraints + strict output shape.
  • Variables named clearly (no vague names like {{data}}).
  • Model chips match real testing.
  • Version + changelog filled before release.

Common mistakes to avoid

  • Too-short descriptions ("does summarization"). Explain scope and format.
  • No negative examples, causing unstable output behavior.
  • Over-tagging every topic, which hurts filter relevance.
  • Publishing as public before validating with real samples.
  • Forgetting to update changelog when editing the body.

The Jonarix CLI requires Node.js 18+. Install it globally with:

npm install -g jonarix

Then set one active API key (stored locally):

jonarix keys set <api_key> jonarix keys status

Verify everything is working:

jonarix --version jonarix keys status jonarix install your-handle/prompt-name jonarix run your-handle/prompt-name

You're ready to connect your studio and run prompts with the same active key.

Your workspace handle is your unique identifier on Jonarix ? similar to a GitHub username. It appears in:

  • Your public profile URL: jonarix.com/your-handle
  • Every prompt you publish: your-handle/prompt-name
  • CLI commands: jonarix run your-handle/prompt-name

It's set during signup and cannot be changed after creation, so choose carefully. Lowercase letters, numbers, and hyphens only (3?32 characters).

Prompts

Quality badges describe how a prompt appears in the catalog. They are separate from visibility (public listing vs only you).

Draft Reviewed Stable
  • Draft ? Not publicly listed (only you), or older prompts that were never switched to the new default.
  • Stable ? Default for prompts that are live in the community catalog. You do not wait for votes or tests to ?earn? this tier before going public.
  • Reviewed ? Reserved for listings the team has manually checked. Pass rates, stars, and saves still help people decide what to run.

Operators can remove policy-breaking prompts from the catalog. If you see abuse, use report flows where available or contact support.

Public prompts appear in the Community catalog, can be starred and installed by anyone, and earn community test results. They contribute to your public profile.

Private prompts are only visible to you. They still get full version control, the same CLI access, and the prompt body is never exposed to other users. Use private mode for:

  • Internal company system prompts
  • Prompts that contain proprietary framing
  • Work in progress before you're ready to share

You can change a prompt from private to public at any time. Going public ? private is also allowed, but it removes the prompt from community listings and search.

Three ways to run any prompt:

1. CLI (recommended):

jonarix run sarah/summarize-article --var text="your article here" --var length=5 --model claude

2. JavaScript SDK:

import { run } from 'jonarix'; const result = await run('sarah/summarize-article', { model: 'gpt-4o', vars: { text: '...', length: 5 } });

3. Python SDK:

from jonarix import run result = run("sarah/summarize-article", model="claude", vars={"text": "...", "length": 5})

All three methods fetch the latest published version unless you pin a specific version with @version, e.g. sarah/summarize-article@1.2.0.

A .jnx file is the portable specification for a Jonarix prompt. It's JSON:

{ "name": "summarize-article", "version": "1.2.0", "description": "Condenses articles into bullet points", "category": "Content", "models": ["gpt-4o", "claude"], "variables": [ { "name": "text", "type": "string", "required": true }, { "name": "length", "type": "number", "default": 5 }, { "name": "tone", "type": "string", "default": "neutral" } ], "prompt": "Summarize the following article in {{length}} bullet points...", "outputSchema": { "type": "list" } }

You can import an existing .jnx file directly: jonarix import ./my-prompt.jnx. Export any prompt with: jonarix export sarah/summarize-article.

Variables

Variables make your prompt reusable. Wrap any dynamic value in double curly braces: {{variable_name}}. When the prompt runs, each {{?}} placeholder is replaced with the real value.

Example prompt body:

Write a {{tone}} email to {{recipient}} about {{topic}}. Keep it under {{word_count}} words.

Run it:

jonarix run you/email-writer \ --var tone="professional" \ --var recipient="the board" \ --var topic="Q3 results" \ --var word_count=150

Variables are auto-detected in the dashboard when you write your prompt body ? just type them and they appear in the Variables section below.

Yes. In the prompt editor, after auto-detection you can mark each variable as required or optional and set a default value for optional ones.

In the .jnx schema:

{ "name": "tone", "type": "string", "default": "professional" } { "name": "length", "type": "number", "default": 5, "min": 1, "max": 20 } { "name": "text", "type": "string", "required": true }

When a caller omits an optional variable, the default is used automatically. If a required variable is missing, the CLI throws an error before sending any tokens.

Models

Jonarix supports any model you have API access to. Current first-class supported models:

  • OpenAI: gpt-4o, gpt-4o-mini, gpt-4-turbo, gpt-3.5-turbo
  • Anthropic: claude-opus-4-7, claude-sonnet-4-6, claude-haiku-4-5
  • Google: gemini-1.5-pro, gemini-1.5-flash, gemini-2.0-flash

When running a prompt, you can pass any model identifier. If the prompt's compatibility list doesn't include it, you'll get a warning but the run will proceed.

jonarix run sarah/code-reviewer --model claude-opus-4-7

When a prompt author publishes, they declare which models they've tested it against. This list is a recommendation, not a hard restriction.

The community test suite then runs the prompt against each listed model and records pass/fail results, token counts, and output quality scores. This is how the Stable badge is earned ? real tests, not self-reported claims.

If you're running a community prompt on an unlisted model, check the Tests tab in the prompt drawer to see how it performed on similar models first.

Versioning

Jonarix uses Semantic Versioning (semver) for prompts: MAJOR.MINOR.PATCH

  • PATCH (1.0.1) ? Minor wording fix, same behaviour and outputs expected.
  • MINOR (1.1.0) ? New optional variable added, backwards-compatible improvement.
  • MAJOR (2.0.0) ? Breaking change: required variable added, output format changed, or intent significantly altered.

Callers can pin a version: jonarix run sarah/summarize@1.2.0, or always get latest: jonarix run sarah/summarize@latest

Every version is stored permanently ? you can diff any two versions in the Versions tab of the prompt drawer.

Yes. Every version is kept forever. You can make any past version the "latest" from the Versions tab, or pin callers to a specific version number.

# Pin to a specific version jonarix run you/my-prompt@1.0.0 # Restore a version as latest (coming in CLI v1.2) jonarix revert you/my-prompt 1.0.0

Teams using jonarix.json can lock prompt versions the same way package.json locks npm dependencies ? preventing unexpected behaviour changes from upstream prompt updates.

CLI Reference
# Installation npm install -g jonarix # Auth jonarix login jonarix logout jonarix whoami # Keys (single active key) jonarix keys set <key> jonarix keys status jonarix keys clear # Prompts jonarix run <handle/name> [--var k=v] [--model m] jonarix add <handle/name> # add to jonarix.json jonarix remove <handle/name> jonarix search <query> # Your prompts jonarix create # interactive wizard jonarix publish <name> jonarix unpublish <name> jonarix version <name> bump [patch|minor|major] jonarix export <handle/name> jonarix import <file.jnx> jonarix diff <handle/name> v1.0.0 v1.1.0 # Agents jonarix agent scan --prompt <handle/name> --path <project_dir> [--model m]

Similar to package.json, jonarix.json is your project's prompt lock file. It records which prompts your project depends on and at what version.

{ "name": "my-ai-app", "prompts": { "sarah/code-reviewer": "^1.4.0", "david/summarize-article": "2.1.0", "priya/data-analyst": "latest" } }

Run jonarix install to fetch all pinned versions. Run jonarix update to upgrade within your semver range. This means your team all runs the same prompt versions ? no surprise behaviour changes.

API Key (BYOK)

Runtime keys are generated by Jonarix and stored in Firestore under your user scope (users/{uid}/apiKeys). Access is restricted to the owning authenticated user (and admin policy where applicable). You can create multiple keys and revoke any key at any time.

Responsibility boundary:

  • Jonarix: key lifecycle management, prompt metadata, workflow orchestration
  • Your runtime environment: CLI/studio session that uses the copied active key
  • Execution backend/model endpoint: validates auth, applies quotas, returns output

Best practice: treat copied keys as credentials with production impact. Avoid sharing in chat logs, repos, or screenshots. Rotate immediately if exposure is suspected.

Billing and hard limits are enforced by the execution provider associated with your runtime path. Jonarix does not override provider quotas or rate limits.

Operational implications:

  • Token spend depends on model selection, input size, output length, and retry behavior
  • 429/limit-type failures usually indicate provider-side rate or quota constraints
  • Permission/scope errors indicate key privilege mismatch for requested endpoint/model
  • Keep install, run, and scan on the same key source to prevent environment drift
  • Use staged rotation to avoid downtime during key replacement

Most integration incidents come from revoked keys, stale local environment variables, or mismatched runtime configuration between terminal and editor extension sessions.

Use a no-downtime rotation workflow:

  1. Generate a new key in Jonarix.
  2. Update CLI/studio to use the new key and run a smoke test.
  3. Revoke the previous key once verification passes.
  4. Record owner/purpose and rotate periodically based on your security policy.

For shared teams, issue distinct keys per environment (local, CI, production automation) to isolate risk and simplify incident response.

Validate in this order:

  • Key status is active in the Jonarix key list
  • Same key is present in both editor and terminal sessions
  • Current shell is using expected env vars (not stale overrides)
  • Requested prompt/model exists and is accessible in your workspace scope
  • CLI version is current and authenticated to the intended account

If needed, generate a fresh key, rebind your tools, and retry with a minimal command path (keys status -> install -> run).

Community catalog

Starring a prompt in the Community catalog saves it under Starred prompts for quick access. Stars are stored locally in your browser and are also synced to Firestore when you're signed in.

The star count on a prompt is a community trust signal ? prompts with more stars tend to be better quality and more widely tested. Stars don't affect stability badges directly, but they do influence which prompts appear in Featured sections.

Yes. Any public prompt can be forked. Forking creates a copy in your workspace that you can modify freely. The fork keeps a reference to the original prompt.

jonarix fork sarah/summarize-article

This creates your-handle/summarize-article as a fork. You can then modify it, publish your improved version, and it'll show "forked from sarah/summarize-article" on its page.

Installing (without forking) just runs the original author's prompt from the registry ? no copy is made in your workspace.

Keyboard Shortcuts
New promptN
Go to DashboardG then O
Go to Your promptsG then P
Go to Community catalogG then D
Go to Developer directoryG then U
Go to Public profileG then B
Go to BYOK API keysG then K
Go to Account settingsG then S
Go to Plans & billingG then L
Go to Help centerG then H
Focus search/
Close drawer / modalEsc
Launch tour?
--var-v
--model-m
--output json-o json
--version--ver
--file-f
--dry-run--dry

Example: jonarix run sarah/code-reviewer -m claude -f main.py -o json

What's New
v0.9.0 May 2026 Features

Light mode, Global Search, Trending Catalog & more

  • Light / dark theme toggle ? saved to your browser across sessions
  • Global omnisearch ? press / and search prompts, @handles, catalog, and settings
  • Trending & Featured sections in the Community Catalog
  • Notifications center ? star alerts, identity review updates, plan limit warnings
  • Prompt forking ? fork any catalog prompt into your own library in one click
  • Import & Export ? download your entire library as JSON or upload .jnx files
  • Collections & Folders ? organize prompts into named groups
  • Comments & discussion on catalog prompts
  • Per-prompt analytics ? runs, stars, and error rate over time
  • Version diff viewer ? compare any two prompt versions side-by-side
  • Visual variable editor ? define variables with types and defaults in the prompt creator
  • In-app prompt playground ? run prompts with your BYOK key directly in the dashboard
  • AI-powered improvement suggestions ? get feedback on your prompts using your own API key
  • Teams workspace UI ? invite members, manage roles, shared prompt library
  • Prompt test suite ? save test cases and replay them after edits
  • Creator tip jar ? community members can tip prompt authors
v0.8.0 Apr 2026 Fixes

3-panel Catalog & Settings sub-tab tracking

  • 3-panel community catalog ? category nav, prompt list, and detail panel
  • Verified tick shown on catalog prompt authors
  • Badge and vertical filter dropdowns removed for cleaner UI
v0.7.0 Mar 2026 Features

Identity verification, 2FA TOTP & public profile

  • Identity verification flow with admin review and purple badge
  • Two-factor authentication with TOTP authenticator app support
  • Public developer profile with README, pinned prompts, and themes
  • Developer directory searchable by @handle, name, and models

Still stuck?

Didn't find what you need? Take the guided tour again, or reach out ? we respond fast.

Create new prompt

Quick mode for fast publishing: add a name, short description, demo media, prompt text, then publish. Switch to Pro for the full advanced workflow.

?

?

No variables defined.
No example output yet.
No prompt recipe available.
No setup instructions available.
No code snippet available.
No build notes available.
No integration playbook available.
0 passed 0 failed
No test results yet. Community tests will appear here once your prompt is published.
# Run directly
jonarix run handle/name
# Add as project dependency
jonarix add handle/name
// In your project
import { run } from 'jonarix';

const result = await run('handle/name', {
  model: 'claude',
  vars: { text: '...' }
});
No issues yet.
No change requests yet.
No activity yet.

Welcome to Jonarix!

You've just signed up for the most powerful AI prompt platform built for developers. Let us take 60 seconds to show you around so you know exactly where everything is.

Version your prompts Track every change, roll back any time, publish to the community.
Community-tested quality Every prompt shows its badge, test results, and real output examples.
BYOK ? your keys stay local API keys never leave your machine. You control the tokens.

You know your way around!

That's everything. You can relaunch the tour any time from the sidebar under Product tour or from Help center. Now ? create your first prompt and start building.