Ship prompts without shipping code
Edit templates, run evaluations, and monitor quality. All in the UI. No PRs, no deploys, no waiting on the release train.
Stop deploying for every prompt tweak. Ship from one dashboard.
Hydra8 keeps your system prompt templates in a single UI. Versioned, testable, and enforced at the gateway. One endpoint in front of every provider. No more config in PRs.
See the difference
Your prompt changes. Your code shouldn't.
Before: Manual
app.ts
// When prompt changes: commit, push, redeploy
// To rollback: another commit, push, redeploy
import { createOpenAI } from '@ai-sdk/openai'
import { generateText } from 'ai'
const SYSTEM_PROMPT = `
You are a helpful assistant. Always respond
in Indonesian. Keep answers concise. Use
examples. Never mention competitors. Format
as JSON with keys: summary, details, action.
If unsure, say so. Be polite. etc etc...
`
const openai = createOpenAI{
apiKey: process.env.OPENAI_API_KEY,
};
const { text } = await generateText{
model: openai('gpt-4o'),
system: SYSTEM_PROMPT,
prompt: 'Hello!',
}
// To rollback: another commit, push, redeploy
import { createOpenAI } from '@ai-sdk/openai'
import { generateText } from 'ai'
const SYSTEM_PROMPT = `
You are a helpful assistant. Always respond
in Indonesian. Keep answers concise. Use
examples. Never mention competitors. Format
as JSON with keys: summary, details, action.
If unsure, say so. Be polite. etc etc...
`
const openai = createOpenAI{
apiKey: process.env.OPENAI_API_KEY,
};
const { text } = await generateText{
model: openai('gpt-4o'),
system: SYSTEM_PROMPT,
prompt: 'Hello!',
}
System prompt hardcoded. Every tweak needs a deploy
No version history. Who changed what, and can you roll back?
Prompt logic lives in code. Not shared with non-devs
After: Hydra8
app.ts
// Change prompt via dashboard. Rollback via dashboard.
// Your code stays clean. Zero deploys.
import { createOpenAICompatible } from '@ai-sdk/openai-compatible'
import { generateText } from 'ai'
const hydra = createOpenAICompatible{
name: 'hydra8',
baseURL: 'https://hydra8/v1',
apiKey: 'sk_your_api_key',
};
const { text } = await generateText{
model: hydra('your-prompt-slug'),
messages: [{ role: 'user', content: 'Hello!' }],
};
// Hydra8 injects the system prompt. Your code stays clean.
// Your code stays clean. Zero deploys.
import { createOpenAICompatible } from '@ai-sdk/openai-compatible'
import { generateText } from 'ai'
const hydra = createOpenAICompatible{
name: 'hydra8',
baseURL: 'https://hydra8/v1',
apiKey: 'sk_your_api_key',
};
const { text } = await generateText{
model: hydra('your-prompt-slug'),
messages: [{ role: 'user', content: 'Hello!' }],
};
// Hydra8 injects the system prompt. Your code stays clean.
Edit prompts in dashboard. No code changes needed
Version history, diffs, rollback. All in the UI
One endpoint. Every provider. Clean code.
Stop writing comparison scripts. Start shipping better prompts.
Your apps call Hydra8. Hydra8 enforces your rules. No PRs, no deploys, no guessing.
Version everything. Promote what works.
Test revisions against live models before you point production at them. Roll back in one click. No redeploy.
Monitor every prompt. Catch failures early.
Set up automated checks against any prompt template. Get alerts when quality drops. No cron jobs, no custom scripts.
One API. Every provider. Open source.
Your apps call Hydra8. Hydra8 routes to OpenAI, Anthropic, or whoever you plug in. Audit the code. Self-host where you need.
The mess you're in
Your prompts live in fifteen repos, three env files, and Dave's Notion.
Every model tweak means a PR. Every guardrail is a one-off validator nobody wrote docs for. Structured output schemas? Copied and pasted across services until nobody knows which one is real.
Your team ships LLM features faster than you can govern them. And you feel it.
How to adopt Hydra8
Connect your providers, write prompts, and ship. No code changes needed.
Connect your providers
Plug in OpenAI, Anthropic, or any OpenAI-compatible provider. One endpoint routes to all of them.
Write and version prompts
Create system prompt templates in the dashboard. Version them, run evaluations, and promote with confidence.
Call Hydra8 from your app
Issue an API key and point your app at Hydra8. The gateway enforces the right template and monitors every call.
Everything you need to ship your features
System prompt templates, evaluations, and monitoring, all in one place. No redeploys needed.
Edit prompts in the dashboard. Changes apply instantly.
Fix typos, tune instructions, and update templates directly in the UI. No PRs, no deploys, no waiting on the release train.
Full version history. One-click rollback.
Every edit is tracked. See diffs, know who changed what, and roll back in a single click. No redeploy needed.
Open source. Self-host. Total control.
Audit every line of code. Run it on your own infrastructure. Plug in any provider without touching your application.
Automated quality checks. Alerts when it matters.
Set up monitors against any prompt template. Hydra8 checks every N minutes, logs scores, and alerts your team when quality drops — all from the dashboard. No cron jobs to configure.
Start free. Scale when you're ready.
Full power on day one. Support when you need it.
We charge for support and fit. Never a markup on model tokens.
Straight answers
What teams ask before they adopt.
- Is Hydra8 actually open source?
- Yes. Every line. Self-host, fork it, audit it. No open-core bait-and-switch.
- How is this different from LiteLLM or Portkey?
- We focus on making prompt management, versioning, and enforcement a first-class UI experience. Not a config file you pray is right. Open source. Self-hostable. Dashboard-first.
- Where do my provider API keys live?
- In Hydra8. Your services get Hydra8-issued keys. Auth and enforcement stay centralized, not scattered across every microservice.
- What happens if Hydra8 goes down?
- You're self-hosting. It's your infrastructure. We don't touch your traffic.
- Can I test a prompt revision before promoting it?
- That's the point. Run test traffic through any revision, compare outputs, then flip production to the one you trust.
Your prompts deserve better than a PR
Ship LLM features without shipping code
Start building with Hydra8 for free. No credit card required.
Try for Free