DevAI
Dashboard
Open source · Bring your own key · Free plan available

One OpenAI-compatible endpoint for many AI providers

Bring your own keys from OpenAI, Anthropic, Gemini, Groq, OpenRouter or Mistral. DevAI routes each request over your keys only, fails over automatically, can race providers in parallel, and shows you every token in one dashboard.

Open dashboard View on GitHub API docs

Works with your keys from

  • OpenAI
  • Anthropic
  • Google Gemini
  • Groq
  • OpenRouter
  • Mistral
  • Ollama Cloud
curl https://devai-gateway.onrender.com/v1/chat/completions \
  -H "Authorization: Bearer dvai_your_key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "auto",
    "messages": [{ "role": "user", "content": "Explain binary search in one line" }],
    "max_tokens": 200
  }'

POST https://devai-gateway.onrender.com/v1/chat/completions

Only on DevAI

Built for students, hackers and free tiers

Unique

Bring your own keys, pooled

Add your own keys from any provider, including free tiers like Gemini, Groq and OpenRouter. DevAI rotates across them, parks a key for a cooldown the moment it hits a 429 and carries on with the next, and shows how far your free quota was stretched. Your requests never touch anyone else’s keys.

Free: 2 keys · Student: 10 · Pro & Team: unlimited

Unique

Hackathon kits

Give every team a temporary API key with a hard request or token cap, an end time and an allowed-model list. Share one link that shows the quota left and the deadline. The gateway cuts off at the cap, and you can revoke a kit at any time.

Included in Team

Features

Everything between your app and the model

Bring your own key

DevAI has no provider keys of its own. You add your OpenAI, Anthropic, Gemini, Groq, OpenRouter or Mistral keys, and your requests run only on them, billed by each provider to you.

Failover + circuit breaker

If a provider errors, times out or rate-limits, DevAI retries the same request on the next one. Failing targets are skipped for a cooldown so users never wait on a dead endpoint.

Many keys per provider

Add several API keys to one provider. A rate-limited or invalid key is cooled down on its own and the request moves to the next key. Choose failover, round robin, random or least used.

Fastest & best parallel modes

fastest sends the prompt to several providers at once and returns the first good answer. best waits for all of them and lets a judge model pick the strongest one.

Response caching

Identical requests are answered from a Redis cache, and identical in-flight requests share one upstream call, saving tokens and latency.

Rate limits & quotas

Per-key token quotas and rate-limit cooldowns keep you inside each provider’s limits and route around keys that are out of budget.

Encrypted provider keys

Provider API keys are stored with AES-256-GCM and never returned in full; the dashboard only shows the prefix and last four characters.

Usage analytics

Every call is recorded with provider, model, tokens, latency, status and cache hits, broken down by provider, model and app.

Per-app keys & routing

Each app gets its own dvai_ key and can have its own routing defaults: mode, provider priority, judge model and timeout.

Streaming

Set stream: true for OpenAI-compatible Server-Sent Events. Failover happens before the first byte, and usage is logged when the stream ends.

Routing

Pick how each request is routed

singleDefault

One provider at a time in priority order. On failure, the next key, then the next provider.

fastestParallel

Ask several providers at once; the first valid answer wins and the rest are cancelled.

bestParallel + judge

Collect every answer, then a judge model picks the most accurate and complete one.

Set a default per app in the dashboard, or override per request with the x-devai-mode header or a devai field in the body.

Quick start

Change one URL. Keep your code.

  1. 1

    Add your provider keys

    Sign in, open Your keys and paste one or more keys from OpenAI, Anthropic, Gemini, Groq, OpenRouter or Mistral. They are encrypted and used only for you.

  2. 2

    Create your DevAI key

    Generate a dvai_ key. Requests with it run only on your keys; with no key added you get a clear no_provider_keys error, never someone else’s quota.

  3. 3

    Call the endpoint

    Use any OpenAI client or plain HTTP with the DevAI base URL and model "auto" or any model your keys support. Streaming works the same way.

curl https://devai-gateway.onrender.com/v1/chat/completions \
  -H "Authorization: Bearer dvai_your_key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "auto",
    "messages": [{ "role": "user", "content": "Explain binary search in one line" }],
    "max_tokens": 200
  }'

Pricing

Simple subscriptions in rupees

Yearly: 2 months free · save 17%

Free

For trying DevAI and side projects.

₹0forever

No card needed

  • 100 requests / day
  • 1 project
  • 2 provider API keys
Get started

Student

College email

For students learning to build with AI.

₹99/month

Billed monthly, cancel anytime

  • 2,000 requests / day
  • 3 projects
  • 5 provider API keys
Popular

Pro

For individual developers shipping apps.

₹499/month

Billed monthly, cancel anytime

  • 20,000 requests / day
  • 10 projects
  • 20 provider API keys

Team

For teams sharing providers, keys and analytics.

₹1,999/month

Billed monthly, cancel anytime

  • 1,00,000 requests / day
  • 50 projects
  • 100 provider API keys

You bring your own provider API keys; provider usage is billed by each provider under your own account.

Compare all features

FAQ

Frequently asked questions

What is DevAI?

DevAI is an AI API gateway: one OpenAI-compatible endpoint that forwards your requests to OpenAI, Anthropic, Gemini, Groq, OpenRouter or Mistral, with automatic failover, key rotation, streaming, caching and usage analytics.

Do I need my own provider API keys?

Yes. DevAI is bring-your-own-key: you add your own provider keys, they are encrypted, and your requests run only on them. Provider usage is billed by each provider to your own account; DevAI never uses anyone else’s quota for you.

Does it work with the OpenAI SDK?

Yes. Point any OpenAI client at the DevAI base URL, use a dvai_ key, and set the model to "auto" or any model your keys support. Chat completions and streaming work the same way.

How much does DevAI cost?

There is a Free plan. Paid plans are subscriptions in Indian rupees through Razorpay, renewing automatically until you cancel: Student ₹99/month, Pro ₹499/month and Team ₹1,999/month, or yearly with two months free (₹990, ₹4,990 and ₹19,990). Cancel anytime; the plan stays until the end of the paid period.

Who can use the Student plan?

Students with a verified college email address (for example .edu or .ac.in). It adds more of your own keys, all key strategies, parallel routing and caching.

What are hackathon kits?

On the Team plan, an organiser can create temporary API keys for hackathon teams, with request and token caps, an expiry time and a share page. Kits run on the organiser’s own provider keys.

Stop wiring every provider by hand

Add your keys once, give each app a dvai_ key, and let DevAI handle failover, parallel routing and usage tracking.

Open dashboard Star on GitHub