Nano GPT logo
NanoGPT
Nano GPT logo
NanoGPT
Conversations
Create Media
Workspace
Gallery
Models
Pricing

Account

Balance
Usage
Settings
API
Presets
Invitations
Subscription
Teams

Resources

Benchmarks
Suggestions and bugs
Batch API
Blog
Help
Applications
Earn
Partners
Terms

|

Privacy

|

Refunds

Private AI

Back to Blog

Tavily Web Search in NanoGPT: BYOK-friendly real-time browsing

Dec 23, 2025
Tavily Web Search in NanoGPT: BYOK-friendly real-time browsing

Web search is the fastest way to keep answers current, but it also adds cost. We added Tavily as a second web search provider so you can keep using NanoGPT while tapping into Tavily's generous free tier via BYOK (bring your own key). The result: more real-time answers for subscribers without extra platform charges.

Why Tavily

Tavily is built specifically for AI web search. It offers a clean credit-based model and a free monthly allotment that makes it a great fit for BYOK users who want fresh information without turning web search into a recurring bill.

From Tavily's docs:

  • Free plan: 1,000 API credits per month, no credit card required.
  • Pay-as-you-go: $0.008 per credit.
  • Search depth: basic searches cost 1 credit, advanced searches cost 2 credits.

(See the Tavily credits and pricing details here: https://docs.tavily.com/documentation/api-credits)

How the integration works

When you enable web search and choose Tavily, we call Tavily's /search endpoint and inject results into the prompt so the model can answer with today's information. Standard search maps to Tavily's basic search depth, while deep search maps to Tavily's advanced search depth.

Pricing inside NanoGPT (and how BYOK changes it)

If you use our hosted Tavily key, web search is billed at:

  • $0.008 per standard search
  • $0.016 per deep search

If you bring your own Tavily key, we skip platform charges for web search. That means subscribers can use their Tavily free tier (1,000 credits/month) on top of their NanoGPT subscription.

How to use Tavily in NanoGPT

In the app

  1. Add your Tavily API key in BYOK settings: https://nano-gpt.com/byok
  2. Go to Settings and select Tavily as your web search provider.
  3. Toggle web search on in your chat and ask a question that benefits from fresh info.

In the API

Append a web search suffix to the model you want to use:

  • Standard Tavily search: :online/tavily
  • Deep Tavily search: :online/tavily-deep

Example model IDs:

  • gpt-4o:online/tavily
  • gpt-4o:online/tavily-deep

When to use standard vs deep

  • Standard: quick checks, headlines, recent releases, fast verification.
  • Deep: research-heavy questions, multi-source synthesis, or when you need more comprehensive coverage.

Why this makes sense

Tavily gives us a second high-quality web search option and gives you control over costs. If you're a subscriber, BYOK lets you add up-to-date web results without extra spend. If you prefer our hosted key, the pricing is still competitive and predictable.

If you want more search options, we still default to Linkup, so you can pick the provider that fits each use case best.

Milan de Reede

Milan de Reede

CEO & Co-Founder

milan@nano-gpt.com

Related articles

Continue with more NanoGPT guides and research on this topic.

How to Choose an AI Web Search Tool

Compare Linkup, Brave, Tavily, Exa, Kagi, Perplexity, Valyu, Sofya, and Firecrawl by cost, search depth, page content, filters, and best use case.

Jul 18, 2026

APIs for Real-Time Text-to-Speech Integration

Compare real-time TTS APIs, solve latency and scaling challenges, and follow best practices for streaming, multilingual voices, and reliable production deployments.

Dec 4, 2025

Real-Time Churn Prediction: Edge Computing Use Cases

Explore how edge computing enhances real-time churn prediction with improved speed, privacy, and cost efficiency for better customer retention.

Sep 6, 2025

When to Use an AI Batch API Instead of Real-Time Requests

Learn when asynchronous AI batch jobs are a better fit than real-time API requests, with practical examples, tradeoffs, and NanoGPT integration steps.

Jul 17, 2026
Back to Blog