With the GigRadar MCP, your AI (Claude, ChatGPT, Gemini and others) can work directly with your GigRadar account/team and investigate or manage quite a lot of things.
š Scanners
List your active scanners
Inspect a scanner's full configuration
Check queries, keywords, exclusions, budgets, countries, client filters, workload, etc.
Preview a scanner and estimate monthly match volume
Search actual jobs matching a scanner/query
Compare overlapping scanners
Analyze scanner priority
Reorder scanners
Duplicate scanners for A/B testing
Create or update scanners
Delete scanners ā only after your confirmation
šÆ Jobs & opportunities
Search the live Upwork job index
Search by technologies, job type, budget, country, client quality, etc.
Inspect a specific opportunity matched by GigRadar
Check its match reasoning and application state
Investigate why a particular job was matched
Analyze whether a scanner is too broad/narrow
š¤ Cover letters / AI behavior
Investigate things like:
Why a particular cover letter was generated
Which information was used for the generation
Whether the job matched the scanner logically
What scanner/template configuration may have influenced it
AI generation/version information when exposed by the data
Match percentage and its reasoning
Problems with Sardor/Laziza-related behavior when the available tools expose the relevant data
For the specific issue we were just investigating, MCP can help me determine which scanner matched the opportunity and what its configuration was. It cannot currently expose every historical cover-letter template/version directly, so DB data like what you pasted can sometimes give us deeper evidence than MCP alone.
š Market research
Analyze Upwork demand:
How many jobs exist for exact niche
Historical job volume
Budget distributions
Client characteristics
Trends for particular niches
Whether a proposed scanner query is likely to produce enough opportunities
š ļø Troubleshooting
Investigate questions such as:
Why did my scanner stop finding jobs?
Why did this job match my scanner?
Why didn't my scanner match this job?
Why was the wrong scanner used?
Why did the autobidder behave differently?
Why was this cover letter generated?
Why is my scanner getting too many irrelevant jobs?
And if something genuinely appears to be a GigRadar bug or missing capability, I can prepare a proper bug/feature report for the GigRadar team rather than just guessing.
ā ļø What AI won't do automatically
Some actions have real consequences, so I'll stop and ask before doing them ā especially:
deleting scanners;
making live scanner changes that could affect bidding;
creating/sending applications or anything that can spend Connects.
So, for example, you can simply tell your AI:
āInvestigate why this job was picked by the React scanner.ā
and it can go into the actual GigRadar data rather than giving you a generic explanation.