> ## Documentation Index
> Fetch the complete documentation index at: https://www.vollna.com/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Recipes

> Prompts for common jobs, and what you get back

Paste a prompt into your connected app and change the names to yours.

## Find why proposals get no replies

**What it does:** finds the funnel stage where your proposals stop and ranks the causes by replies lost.

**Tools used:** `diagnose_reply_rate`, `get_proposal_funnel`

**Prompt:**

```
Why do my proposals get so few replies? Use the last 30 days and rank the causes.
```

**What to expect:** the stage where proposals stop, the share of jobs where the client hired nobody, and each cause with the replies it cost. Nothing changes.

***

## Clean up a noisy filter

**What it does:** finds jobs that don't fit, explains why each matched, and tests one criteria change before saving it.

**Tools used:** `get_filter_jobs`, `explain_job_in_filter`, `preview_filter_changes`, `update_filter`

**Prompt:**

```
Show the jobs from the last 7 days in my "Shopify agencies" filter that don't fit a Shopify developer. For each, say why it matched, then propose one criteria change that keeps them out.
```

**What to expect:** the jobs that don't fit, the keyword or setting that let each in, and how many jobs the filter matches before and after the change over 30 days. The filter changes only when you ask the app to save.

<Note>
  Every job the app reads uses one project credit, the first time in a usage period.
</Note>

***

## Test a qualifier prompt before you save it

**What it does:** runs a new qualifier prompt on past jobs without saving it, and shows which verdicts flip.

**Tools used:** `replay_qualifier_prompt`, `save_filter_qualifier_prompt`

**Prompt:**

```
Add a rule to the qualifier of my "Laravel" filter: reject jobs under $500. Replay it on 10 qualified and 10 disqualified jobs and show which verdicts flip.
```

**What to expect:** each job with its old and new verdict, and a count of flips. When it looks right, say "Save it". The app asks you first when other filters share the qualifier.

<Warning>
  A replay spends AI qualifier credits, one per job.
</Warning>

***

## Find a niche worth bidding on

**What it does:** compares competition, rates and demand between niches.

**Tools used:** `get_market_insights`, `get_niche_difficulty`, `get_keyword_opportunities`

**Prompt:**

```
Compare competition and rates for React and Vue jobs over the last 90 days. Which one gives me better odds?
```

**What to expect:** applicants per job, interview and hire rates, a difficulty score from 0 to 100, and the keywords freelancers in each niche track. Nothing changes. Each of these reads has a [daily cap](/docs/mcp/changes-and-limits#daily-caps).

***

## Improve a proposal template

**What it does:** finds your weakest template, compares its best and worst letters, and tests one change on real jobs.

**Tools used:** `get_template_stats`, `get_template_report`, `preview_proposal`, `update_proposal_template`

**Prompt:**

```
Which of my proposal templates gets the fewest replies? Compare its best and worst letters and suggest one change to the instructions. Preview the new letter on two recent jobs.
```

**What to expect:** replies per template, what the replied letters do that the ignored ones don't, and two test letters with the change. The template changes only when you ask. The app asks you first when an active auto-bidding setup sends with it.

<Note>
  Each test letter uses one from your AI proposal preview quota.
</Note>

***

## Check auto-bidding

**What it does:** shows what stops each setup from bidding and where connects went.

**Tools used:** `autobidding_status`, `autobidding_list_setups`, `autobidding_analytics`

**Prompt:**

```
Check my auto-bidding setups. What stops each one from bidding, and how many connects did each spend this month?
```

**What to expect:** each setup with what blocks it, and connects spent per setup. Nothing changes. Pausing or activating a setup always asks you first.

***

## Weekly summary

**What it does:** sums up the last 7 days across filters, proposals and auto-bidding.

**Tools used:** `list_filters`, `get_proposal_funnel`, `autobidding_analytics`

**Prompt:**

```
Summarize my last 7 days: jobs matched per filter, proposals sent, replies, connects spent. What should I change this week?
```

**What to expect:** a table of the week and the changes the app suggests. Nothing changes until you ask.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.