Orchex Beta Promise

Version: 3.0 Last Updated: 2026-02-23 Status: Beta ready (multi-LLM pivot)

What We Promise Beta Users

The Core Promise

Turn a plan into parallel AI agents that can't break each other's code.

Paste a plan doc into your AI assistant. Orchex splits it into parallel streams — each locked to its own files, each self-healing on failure. 5 LLM providers, dependency-aware waves, ownership enforcement. What took 2 hours of serial prompting becomes 12 minutes.

What You Get

Feature-Gated Local (Free)

• Ownership enforcement — Streams can only modify their declared files
• Parallel execution — Up to 5 streams, 2 waves
• Single provider — One LLM at a time
• Manual retry — Re-run failed streams (self-healing requires Pro+)
• BYOK — Use your own API key
• Works offline — No account needed for local use

Cloud Trial ($5 Credit, 30 Days)

• Full cloud features — Unlimited streams/waves, all 5 providers
• orchex learn — Markdown plan → parallel executable streams
• Observability dashboard — Stream timelines, diffs, cost tracking
• No credit card required — $5 credit to experience cloud execution
• Direct developer access — Feedback goes straight to the builder
• Shape the roadmap — Your use cases influence what gets built

What We Ask

Minimum (Required)

1. Try it on a real project — Not a toy example, something you actually want to build
2. Report bugs — Email support@orchex.dev or use the in-app feedback form

Appreciated (Optional)

3. Share your use case — What are you building? How does orchex help?
4. Documentation feedback — What was confusing? What's missing?
5. Spread the word — If it works for you, tell others

What We Don't Promise

Not Yet

• Enterprise features — SSO, SCIM, audit logs for compliance
• Guaranteed uptime — Beta means things will break
• 24/7 support — Solo developer, response times vary

Never

• Magic — You still need to think about your architecture
• Replacement for understanding — Orchex executes your plan, it doesn't make the plan
• Autonomous decision-making — You define streams, you review results
• File modifications outside ownership — Streams can only modify their declared files

The Beta Journey

Closed Beta Criteria

• Invite-only (signup approval required)
• Target: 80%+ complete onboarding checklist
• Focus: First orchestration success, documentation feedback

Open Beta Criteria

• Public signup enabled
• Self-serve onboarding
• Focus: Can you succeed without hand-holding?

Who This Is For

Perfect Fit

• OpenAI and Gemini users wanting parallel orchestration
• Developers wanting provider flexibility — not locked to one LLM
• Teams needing file safety — ownership enforcement prevents accidents
• Solo developers building features with AI assistance
• Small teams (2-5) shipping faster than competitors
• Local LLM users running Ollama for privacy

Good Fit

• Anyone using GPT, Gemini, Claude, or local models
• Developers frustrated by unreliable AI agents
• Teams wanting consistent orchestration across different LLMs

Not Yet

• Enterprises needing regulatory compliance certifications
• Large teams (20+) with complex coordination needs
• Organizations requiring managed API keys

Not Targeting

• Developers working on small tasks (1-3 files) — serial execution is fine for small work

The Problem We Solve

Before Orchex

You: "GPT, update the types file"
[Wait 2 min]
You: "Now update the API file"
[Wait 2 min — oh no, it modified the config file too]
You: "Now update the tests"
[Wait 2 min — test failure, manual retry]
You: "Now update the docs"
[Wait 2 min]

Total: 8+ minutes of waiting, files modified unexpectedly, manual error handling

After Orchex

You: Define 4 streams (types, api, tests, docs) with ownership
Orchex: Runs all 4 in parallel with your LLM
Orchex: Each stream can only modify its declared files
Orchex: Test failure → categorizes error → auto-generates fix stream

Total: ~2 minutes, file safety guaranteed, self-healing built-in

The Transformation

Feedback Channels

Bug Reports

• Email: support@orchex.dev
• Include: What you did, what happened, what you expected

Feature Requests

• Email: support@orchex.dev with subject "Feature Request"

Direct Feedback

• Email: founders@orchex.dev
• For: Use cases, general feedback, partnership inquiries

Our Commitment to You

1. Transparency — We'll be honest about what works and what doesn't
2. Responsiveness — Bug reports get attention (may not be instant, but they're seen)
3. Iteration — Your feedback shapes the product
4. Respect — Your time matters; we won't waste it with hype

FAQ

What AI providers are supported?

OpenAI (GPT-4, GPT-4.1), Google Gemini (1.5, 2.0), Anthropic Claude, DeepSeek (V3, Coder, Reasoner), and local models via Ollama. Configure your provider with a single environment variable.

Is it really free?

Local use: Feature-gated free. Max 5 streams, 2 waves, 1 provider. Enough to try, not enough for real workloads. Cloud trial: $5 one-time credit, 30 days. Full features, no credit card. Paid tiers: Pro $19/mo (100 runs), Team $49/user/mo (500 runs), Enterprise custom.

What is ownership enforcement?

Each stream declares which files it can modify in its owns array. Orchex rejects any file operations outside that list. This prevents agents from accidentally modifying files they shouldn't touch.

Do you store my code?

Local: No code leaves your machine. Cloud: Context sent for LLM calls, not persisted on our servers.

What if orchex breaks my code?

Orchex creates backups before applying changes. Ownership enforcement prevents most accidental modifications. You can always rollback.

Can I use this at work?

Yes, subject to your company's policies on AI tools. BYOK means you control API spend with your preferred provider.

What happens after beta?

Feature-gated local stays free. Cloud pricing is published. No bait-and-switch. Your $5 trial credit will be honored.

End of document.

This is our promise. We intend to keep it.