Free API Keys (NaraRouter)

NaraRouter is a unified AI model gateway — one API key gives you access to 40+ models through a single OpenAI-compatible endpoint. Its free tier is generous enough to run an agent on.

• Website: https://router.bynara.id/
• Sign up: https://router.bynara.id/register
• Model list: https://router.bynara.id/ai-models
• Docs: https://router.bynara.id/docs

Sign up — step by step

1. Go to https://router.bynara.id/register.
2. Sign up with email + password (min 8 characters) or Continue with Google. No credit card required.
3. Link your Telegram account in the dashboard — this is required before the /v1 endpoints will answer. If your calls return 401/403 right after signup, this is almost always the cause.
4. Open the dashboard → API Keys → Create key.
5. Keys start with sk-nry- and are shown only once — copy it immediately. You can rotate or revoke it later from the dashboard.

Test your key

curl -s -X POST https://router.bynara.id/v1/chat/completions \
  -H "Authorization: Bearer sk-nry-YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"agnes-2.5-flash","messages":[{"role":"user","content":"hi"}],"max_tokens":10}'

A working key returns a JSON reply. To see exactly which models your plan includes:

curl -s https://router.bynara.id/v1/models \
  -H "Authorization: Bearer sk-nry-YOUR_KEY"

Free-tier models

Model ID	Notes
agnes-2.5-flash	Best free default — fastest, reliable, vision-capable
agnes-2.0-flash	Free tier, older generation
tencent-hy3	Strong reasoning model
Mistral Large (free listing)	Marked "Free" in the model directory
Kimi K2.7 Code Free	Free coding model
Poolside / NVIDIA models	Listed at Rp0/$0 (free)

Not free (need credits): deepseek-v4-pro-0813-bynara, Anthropic/OpenAI flagship IDs. Note: deepseek-v4-flash-free has a misleading name — it's not on the free tier.

Free-tier limits

• ~6M tokens/day, resets 07:00 WIB (00:00 UTC)
• 10–15 requests/minute
• Free responses can take 5–20s on shared infrastructure — normal.

Values to paste into AgentX

Field	Value
Base URL	https://router.bynara.id/v1
API key	Your sk-nry-… key

The app uses the standard OpenAI chat-completions format, which NaraRouter speaks natively.

Alternatives

• OpenRouter (openrouter.ai) — one key, free models with the :free suffix (e.g. tencent/hy3:free). Base URL https://openrouter.ai/api/v1. 20 req/min, 50 req/day free (1,000/day after a one-time $10 credit purchase). Good fallback lane.
• Google AI Studio (aistudio.google.com) — ~1,500 req/day free, direct provider, no card. The most trustworthy free option.
• Groq (console.groq.com) — 30 RPM / ~14.4K req/day free, very fast inference. No card.

Treat resellers as a bonus, not your only key

NaraRouter is an aggregator, not an inference provider. Keep a second provider (e.g. Google AI Studio) configured so one outage doesn't strand you.

What's next

• Add a Provider — where the base URL and key go in the app.
• Choose Models — which model for which job.