Make ClawdBot AI Learn to Search in 5 Minutes: This Free API is Amazing
Hello everyone, I'm Xiaoxiao.
I introduced Moltbot (formerly Clawdbot) to you before, the AI tool that can control your computer through Telegram. I've been playing with it since I installed it, and I found a problem: It can't search the internet.
For example, if I ask it "What's the latest AI news?", it can only answer based on the training data, and the information is outdated. This is like asking someone who has only read books from 2023 to tell you today's news, how can it be accurate?
Later, I flipped through the documentation and found that Clawdbot supports configuring a search engine, and the official recommendation is Brave Search. Today, I'll share my configuration process with you, it only takes 5 minutes, and the API is free.
In fact, there are many search APIs available on the market, including Google and Bing, but I finally chose Brave for a simple reason:
- Sufficient Free Quota: Google API offers 100 free requests per day, Bing offers 1000 per month, and Brave offers 2000 per month. For personal testing, 2000 requests are really enough.
- Better Privacy Protection: Brave focuses on privacy and doesn't track users, which I appreciate. As a tool enthusiast, privacy is always my bottom line.
- Simple Configuration: Unlike Google API, which requires opening a bunch of services and configuring OAuth, Brave uses simple API key authentication, one step at a time.
- Good Search Quality: I tested a dozen queries, and the result quality was better than I expected, at least it's fine for daily use.
Before starting the tutorial, let me mention a pitfall I encountered: Brave registration API requires VISA card verification**.
Its payment system uses Stripe, and it will deduct $1 for verification and then refund it immediately. This is to verify that you are not a robot, but this operation does deter many people.
This step is actually very simple, just follow the instructions.
1. Register an Account
Visit the Brave Search API official website: https://brave.com/search/api/
Click "Get Started" to register an account. Fill in your email address, use your frequently used email address, as you will need it to verify the API key later.
2. Bind a VISA Card
You will be prompted to bind a bank card when subscribing, just follow the process. The deducted $1 will be refunded within a few minutes, so don't worry.
3. Subscribe to the Free Plan and Create an API Key
After logging in, you will see the console, subscribe to the free plan.

Then click the "Add API Key" button.
Here are a few options:
- API Key Name: Fill in anything, such as "Clawdbot"
- Use Case: Select "Personal use"
- Quota: The free version offers 2000 calls per month
After clicking create, the system will give you a string of characters, which is your API Key.
After getting the API Key, there are two configuration methods. I've tried them all, and I'll tell you the differences.
This is the official recommended method, and it's also the one I use most often:
clawdbot configure --section web
After execution, you will enter interactive configuration, prompting you to enter the Brave API key:
? Enter your Brave Search API key: [Paste your key]
Just paste the key you just copied, and press Enter to complete it.
Advantages:
- Simple configuration, done with one command
- Automatically saved to the configuration file
- Supports hot updates, no need to restart
I used this method at the time, and it took less than 30 seconds from execution to configuration.
If you prefer manual configuration, or need to use it in special environments such as Docker, you can set environment variables:
# Add to ~/.bashrc or ~/.zshrc
export BRAVE_API_KEY="your key"
# Reload the configuration
source ~/.bashrc
Note: You need to restart the Gateway for this configuration to take effect.
Restart command:
clawdbot gateway restart
After the configuration is complete, you can test it.
Test via Telegram
Send a message to your Clawdbot:
Search: the latest AI technology trends in 2026
After a few seconds, it will return the search results, including:
- The title and link of the relevant webpage
- A brief summary
- Source website
I was really shocked when I first saw this feature - AI can finally get information in real time.
Command Line Test
If you like the command line, you can also test directly:
clawdbot search "the latest AI news"
This command will directly output the search results, which is convenient for debugging.
After configuring Brave search, I used it for a few days, and I'll share my real feelings.
Good Use Cases
- Quickly Obtain Information: "Search: What are the major events in the tech circle today?" It will give you the latest news list, which is much faster than browsing tech media yourself.
- Technical Problem Query**: "Search: How does Python asynchronously handle HTTP requests?" It will search technical communities such as Stack Overflow and summarize the solutions.
- Data Collection**: "Search: AI entrepreneurship directions in 2026" Suitable for quickly collecting information when doing research.
Not Suitable Scenarios
- Complex Queries: For example, "Compare the advantages and disadvantages of iPhone and Android", the search effect is average for this kind of problem that requires in-depth analysis.
- High Real-Time Requirements: Although it can connect to the Internet, it takes a few seconds to ten seconds for searching + understanding + replying, and there will be a delay in the chat scenario.
- Need for Data Accuracy: For example, stock prices and weather information, it is recommended to use a dedicated API instead of relying on search.
Configure Brave search for Clawdbot,
It takes 5 minutes, the free quota is enough, and the search quality is good. Most importantly, it turns AI from "just chatting" into "a working information assistant".
If you are also using Clawdbot or Moltbot, I strongly recommend spending 5 minutes configuring it. Especially when doing technical research and data collection, this function can really save a lot of time.
Finally, a reminder: keep your API key safe**. Although it is just a search API, if it is leaked, it may be abused, causing your quota to be exhausted.




