---
name: trending-jobs
description: "Step-by-step job recipes for the Trending Now MCP server: Rising posts for a topic; Daily trend digest for a niche; Open deals and freebies. Use when the user asks things like \"What is trending in AI in the last 6 hours?\" or \"Give me today's content ideas digest for marketing.\""
license: MIT
---

# Trending Now jobs

Each section below walks through one job with the tools of the `trending` MCP server. The same recipes are MCP prompts (prompts/list) on `https://trending.openkrill.app/mcp`.
Connect the server at `https://trending.openkrill.app/mcp` in your MCP client settings (remote streamable HTTP, no key), then call its tools as normal tool calls. They are not command line programs: Bash, curl and scripts cannot reach them.

## Rising posts for a topic

Job: Rising posts for a topic, with the Trending Now MCP server.

Use the `trending` tool for this job.

Work on the user's request. If there is none yet, ask what they need. A request this job handles: "What is trending in AI in the last 6 hours?"

Steps:
1. Call `trending` as a normal MCP tool call, with the request's values in the fields of its inputSchema.
2. If a value the tool needs is missing and you cannot tell it from the request, ask the user one short question first.
3. This input is known to work; copy its format, not its values: `{}`
4. For the request "What is trending in AI in the last 6 hours?", a correct answer looks like this: AI posts ranked by heat, each with link, source, count, unit and gain per hour, plus rising keywords.
5. Follow the server's own rules:
   - Call trending when the user wants what is popular or rising now.
   - Pass topic as a niche slug or the user's own words, and window (1h, 6h, 24h, 3d, 7d; default 24h).
   - Give each post's title, link, source, and its count with unit and velocity_per_hour.
   - Say heat compares a post only with others from the same source.
   - For a top list, take posts in rank order and say the rule: rank orders every post by heat, highest first; posts with equal heat (each source's top post has 100) go newer first. velocity_basis since_last_read is measured between two reads; since_published is the average since posting; list_order means the source gives no count (Product Hunt).
   - Quote rising_keywords with change_percent only when keywords_baseline is ready; while it is building, give the post counts and say the comparison needs six days of data.
   - If status is warming, say the first read is running and to ask again in a minute.
   - If no_posts, suggest a longer window or a wider topic.
   - Name any source whose status is failed.
   - Reddit and X are not read: say so from not_read and never invent posts from them.
   - Reddit is not read: its Data API terms need a separate agreement for commercial use.
   - Answers hold titles and links only; do not invent post text, authors or counts.
6. If the call fails or the answer looks wrong, tell the user, then report it with `submit_feedback` (kind bug, tool `trending`).

## Daily trend digest for a niche

Job: Daily trend digest for a niche, with the Trending Now MCP server.

Use the `trending_digest` tool for this job.

Work on the user's request. If there is none yet, ask what they need. A request this job handles: "Give me today's content ideas digest for marketing."

Steps:
1. Call `trending_digest` as a normal MCP tool call, with the request's values in the fields of its inputSchema.
2. If a value the tool needs is missing and you cannot tell it from the request, ask the user one short question first.
3. This input is known to work; copy its format, not its values: `{"niche":"developer-tools"}`
4. For the request "Give me today's content ideas digest for marketing.", a correct answer looks like this: Top marketing posts, the top 5 per source, rising keywords and the page link.
5. Follow the server's own rules:
   - Pass topic as a niche slug or the user's own words, and window (1h, 6h, 24h, 3d, 7d; default 24h).
   - Give each post's title, link, source, and its count with unit and velocity_per_hour.
   - Say heat compares a post only with others from the same source.
   - For a top list, take posts in rank order and say the rule: rank orders every post by heat, highest first; posts with equal heat (each source's top post has 100) go newer first. velocity_basis since_last_read is measured between two reads; since_published is the average since posting; list_order means the source gives no count (Product Hunt).
   - Quote rising_keywords with change_percent only when keywords_baseline is ready; while it is building, give the post counts and say the comparison needs six days of data.
   - Call trending_digest for a niche's daily digest or content ideas, and give its page link.
   - If status is warming, say the first read is running and to ask again in a minute.
   - If no_posts, suggest a longer window or a wider topic.
   - Name any source whose status is failed.
   - Reddit and X are not read: say so from not_read and never invent posts from them.
   - Reddit is not read: its Data API terms need a separate agreement for commercial use.
   - Answers hold titles and links only; do not invent post text, authors or counts.
6. If the call fails or the answer looks wrong, tell the user, then report it with `submit_feedback` (kind bug, tool `trending_digest`).

## Open deals and freebies

Job: Open deals and freebies, with the Trending Now MCP server.

Use the `trending_deals` tool for this job.

Work on the user's request. If there is none yet, ask what they need. A request this job handles: "Any free Steam games I can claim right now?"

Steps:
1. Call `trending_deals` as a normal MCP tool call, with the request's values in the fields of its inputSchema.
2. If a value the tool needs is missing and you cannot tell it from the request, ask the user one short question first.
3. This input is known to work; copy its format, not its values: `{"query":"steam key"}`
4. For the request "Any free Steam games I can claim right now?", a correct answer looks like this: Open Steam giveaways, newest first, each with its link and free_until when the source gives an end date, plus the GamerPower credit.
5. Follow the server's own rules:
   - Pass topic as a niche slug or the user's own words, and window (1h, 6h, 24h, 3d, 7d; default 24h).
   - Give each post's title, link, source, and its count with unit and velocity_per_hour.
   - Say heat compares a post only with others from the same source.
   - For a top list, take posts in rank order and say the rule: rank orders every post by heat, highest first; posts with equal heat (each source's top post has 100) go newer first. velocity_basis since_last_read is measured between two reads; since_published is the average since posting; list_order means the source gives no count (Product Hunt).
   - Quote rising_keywords with change_percent only when keywords_baseline is ready; while it is building, give the post counts and say the comparison needs six days of data.
   - If status is warming, say the first read is running and to ask again in a minute.
   - If no_posts, suggest a longer window or a wider topic.
   - Name any source whose status is failed.
   - Reddit and X are not read: say so from not_read and never invent posts from them.
   - Reddit is not read: its Data API terms need a separate agreement for commercial use.
   - Answers hold titles and links only; do not invent post text, authors or counts.
6. If the call fails or the answer looks wrong, tell the user, then report it with `submit_feedback` (kind bug, tool `trending_deals`).