How to use Jina DeepSearch with BoltAI
Last updated
Was this helpful?
Last updated
Was this helpful?
Jina DeepSearch is an “agent-style” API endpoint that loops through search → read → reason → search until it finds the most complete answer to your prompt, then streams the result (and its thought process in tags) back to you. It is built on Jina’s Reader LM-v2 plus web-search utilities and optimised for high recall on the very first answer rather than for long-form generation.
It's easy to set up Jina DeepSearch in BoltAI.
Get a . Each new key starts with 1 million free token.
Add the key in BoltAI.
Start using Jina DeepSearch in a new chat.
You need a Jina API Key to use DeepSearch.
Sign up for .
In Jina dashboard, click "Add key" (1), then click "Generate new key" (2) in the dialog.
You should see a new API key in the list.
Go to Settings > Services, click the plus button and select Jina DeepSearch. Enter your API key and click Save Changes.
Jina DeepSearch works similarly to any other AI service provider. You can select it from the provider dropdown and then start the chat.
Note that it might take a few minutes for each deep search request. Make sure to increase the network timeout settings accordingly. You can find the network timeout setting in Settings > Advanced > Network.
How is DeepSearch different from OpenAI and Gemini's deep research capabilities? Unlike OpenAI and Gemini, DeepSearch specifically focuses on delivering accurate answers through iteration rather than generating long-form articles. It's optimized for quick, precise answers from deep web search rather than creating comprehensive reports.
Why does DeepSearch use so many tokens for my queries? Yes, the token usage of DeepSearch on complex queries is arguably high - averaging 70,000 tokens compared to 500 for basic LLM responses. This shows the depth of research but also has cost implications.
Is there a way to control or limit the number of steps? The system is primarily controlled by token budget rather than step count. Once the token budget is exceeded, it enters Beast Mode for final answer generation. Check reasoning_effort for more details.
More questions Learn more from
