GitHubNew Term Radar

Run New Term Radar with AIsa

Data and models for New Term Radar, through AIsaX / TwitterAccount updates · keyword searchDataForSEOSearch volume · trends · new domains
GPTClaudeModel support

New-term identification · keyword selection

01 / NEW-TERM RADAR

Set up and start using New Term Radar

Read the repository ↗

What is New Term Radar?

New Term Radar is an open-source project that tracks emerging AI terms across X, Hugging Face, GitHub and website sitemaps. It compares new signals with saved history, uses AIsa’s data and models to assess terms and search demand, and brings the findings into a daily report.

Claude CodeCursor

Let your agent install New Term Radar

Use Claude Code, Cursor, or another agent that can create files and run Python. Enter the actual key locally when it guides you.

Project setup prompt
Help me install and configure New Term Radar locally.
Repository: https://github.com/zhuyansen/new-term-radar
Read the README, check dependencies, and guide me through sources and AIsa key setup.
…
Run it on a schedule with GitHub Actions

Fork the repo. Remove config.json from .gitignore and commit your non-secret configuration. Add AISA_API_KEY under Actions secrets, enable workflows, and set the repository variable RADAR_ENABLED to true. Follow the repository instructions for optional services and state backups.

Full scheduling instructions ↗

02 / YOUR AGENT

Connect your agent and start researching with AIsa

You can also connect AIsa directly to your agent for a research task of your own. Choose your topics and sources, then let your agent gather and organize the findings.

Connect your agent

  1. 1

    Add the AIsa plugin

    AIsa is in ChatGPT's plugin directory. Open it and add it to ChatGPT:

    Open AIsa in ChatGPT
  2. 2

    Sign in with OAuth

    ChatGPT opens a browser sign-in. Use your AIsa account. There's no API key to paste.

  3. 3

    Create your own prompt

    Go to the prompt builder below, customize your task, then copy the finished prompt into ChatGPT.

    Customize my prompt

Choose what your agent should research

Choose your research scope and send the prompt to your agent after connecting AIsa.

Read and edit your research prompt
Research prompt
Use AIsa to find emerging topics worth researching for my next content page.
Niche: AI tools for independent creators
Market and language: United States · English
Starting sources: Public X posts, GitHub projects and product websites
API budget: Up to $1; ask before spending

First check available tools and pricing, propose a bounded plan, and ask before paid calls. Find recent products, projects and recurring user problems in the sources you can access. Separate new names from generic tasks. For each promising lead, verify the original source and publication date, check regional search demand and trend context where available, and inspect existing search results. Do not claim a term is new without historical evidence.
Return a shortlist with source links, the user problem, candidate search query, evidence and gaps, and a decision: research now, watch, or skip. Recommend one page idea, its audience, and an outline. Label estimates and missing data; do not infer traffic or sales from social engagement. This is a one-off research task, not a request to install software or schedule monitoring.

A few practical questions

What is New Term Radar?

New Term Radar is an open-source project that tracks emerging AI terms across X, Hugging Face, GitHub and website sitemaps. It compares new signals with saved history, uses AIsa’s data and models to assess terms and search demand, and brings the findings into a daily report.

Where do I get my API key?

Click Get API Key, sign up or sign in, then create your key in the console. An existing AIsa key works too.

Do I need the repository?

Download it for the ready-made radar and its saved history. For ad hoc research, connect AIsa directly to your agent instead.

Why is my first report short?

The first run establishes a baseline. Later runs compare changes; a new mention is a lead to investigate, not proof of future traffic.

What should I do if a source fails?

Check the repository’s current Known issues. The author reports an issue with the new-sites signal; --skip newsites lets you leave that signal out. An error is not evidence of zero demand.

How much does it cost?

API usage depends on the sources, models and frequency you choose. Ask your agent to confirm the available tools and estimated cost before it starts. Running more accounts or more frequent checks increases usage.