TGGET Niche Research

What TGGET Niche Research does

TGGET Niche Research MCP server gives an AI assistant access to TGGET’s hosted niche research service at https://tgget.io/mcp. It is intended for checking mobile app and online service ideas against measured search and market data rather than relying only on a written concept.

The TGGET Niche Research MCP server accepts either a search phrase or a plain-language description. A phrase is researched directly; a description is converted into likely search queries and the strongest candidate is scanned. Results can include monthly demand, related phrases, competitors in search results, App Store apps and reviews, demand history, paid search metrics, and public interest signals.

How it works

Research runs asynchronously. Call start_research to receive an id, then poll get_research every 20–30 seconds until the status is completed and no work remains pending. Use get_report to retrieve the finished report in Markdown or structured JSON.

Demand is measured for one country and language. If these are not supplied, TGGET selects them from the phrase, using the corresponding market for English, Spanish, or Portuguese. The returned country and language identify the market behind the measurements.

The service can compare two to five completed scans with compare_research. It can also follow a completed niche with follow_niche; monthly checks use the same market and provide changes from the previous measurement. Project history is available through list_projects and get_project.

Setup and configuration

TGGET Niche Research is a remote Streamable HTTP server. Connect to https://tgget.io/mcp and send a personal access token as a Bearer token. The endpoint can describe itself and list tools without authentication, but tool calls require a token.

Create an account at TGGET or in its Android app, confirm the email address, and create a token in Settings under “API and AI assistants.” Tokens are displayed once, expire after one year, and can be revoked by changing the password or closing sessions. Cursor, VS Code, Claude Code, and Codex connection examples are provided in the repository. Codex can read the token from the TGGET_TOKEN environment variable.

Tools and capabilities

The TGGET Niche Research MCP server exposes tools for starting, polling, listing, and reporting research:

• start_research, get_research, get_report, and list_research manage scans and their results.
• run_signal adds checks such as competitors, dynamics, App Store data, reviews, community signals, paid search, insights, or a rebuilt report. These added checks do not consume research allowance.
• compare_research ranks measures across finished reports using their numeric values.
• follow_niche, list_projects, get_project, and update_project manage recurring monthly monitoring.

The server also provides the tgget://quota and tgget://sources resources, plus a find_niche prompt that guides an assistant through starting research, waiting for completion, reviewing evidence, and writing a verdict.

Limitations and notes

A scan counts against the plan when it starts, including when data comes from shared cache. The free plan provides three initial researches and one additional research every 30 days. Followed niches are limited to one on Free, two on Starter, and ten on Pro. Monthly checks do not consume research credits.

Free-plan research is public and may appear in TGGET’s niche catalog; the typed description is not shown there. The API and MCP endpoint share a limit of 60 requests per minute. Search interest is evidence of demand, not a customer count, so the verdict should be treated as an input to further validation. Demand and cost-per-click comparisons are not ranked across different countries or languages.