# AI Apps That Work free tools (remote · aiappsthatwork.com)

Real cost per AI result, LLM prices, switch costs, shutdown dates, A/B test sizes and judge checks.

- Trust score: 75/100 (medium)
- Change this week: +3
- Registry status: active
- Liveness: live
- Owner verified: no
- Last scored: 2026-11-21

## Components

- remote · `aiappsthatwork.com`: 75/100 (this document), [markdown](https://verifymcp.io/servers/com-aiappsthatwork-tools/aiappsthatwork.md), [page](https://verifymcp.io/servers/com-aiappsthatwork-tools/aiappsthatwork)

## Channel facts

- Endpoint: `https://aiappsthatwork.com/mcp`
- Transports: `streamable-http`
- Auth: `none`
- Version: `1.0.1`

## Trust breakdown

How this component scores in each security and reliability category. Every signal is checked automatically against the live server, and we only credit what we can confirm. Scores are 0–100 per category. Scoring method: https://verifymcp.io/docs/scoring (what has changed: https://verifymcp.io/docs/scoring/changelog)

Scored 2026-11-21.

- **Endpoint Security**: 80/100
  - The endpoint's TLS certificate is valid, in date, and uses a strong key.
  - No authorisation is required to call this server. Nothing on it is marked destructive, but we can't show that an anonymous caller can only read: full credit needs the server to list at least one tool and every tool to be verifiably read-only. We use a best-effort check of tool names and descriptions to tell read-only tools from tools that change things, so some genuinely read-only tools may not get full credit. Names that start with a read verb (get, list, read, search and similar) and accurate readOnlyHint annotations help. Meet that to earn full credit, or require authorisation.
  - HTTPS is enforced; there's no plaintext access path.
  - The HSTS (Strict-Transport-Security) header is present.
  - DNSSEC check failed: this domain isn't protected by DNSSEC.
- **Transport & Reachability**: 100/100
  - Verified streamable-http transport via a live MCP handshake.
- **Schema Quality & AI Usability**: 60/100
  - AI-judged instruction clarity (good).
  - Context-footprint check failed: tool/resource definitions use about 1885 tokens (~235/item across 8 items; 8 tools + 0 resources), over budget; trim descriptions and params.
  - Usage-examples check failed: none of the tools include examples.
- **Stability & Change Management**: 30/100
  - Stability observed for 9 of 30 days with no destabilising changes; credit accrues until the full window elapses.
- **Tool Coverage**: 100/100
  - 100% of tools have a non-trivial description (not blank, and not just the tool's name).
  - 100% of tool parameters carry a description.
  - Structured output schemas are declared (100% of tools); any adoption earns full credit.
- **Tool Safety**: 100/100
  - No prompt-injection markers were found in the server instructions, tool names or descriptions we captured.
  - We read all 8 captured tool definition(s), and no name or description among them implies an irreversible operation.
  - An AI judge read all 9 captured unit(s) of tool text and found none that tries to manipulate the model reading it.
- **Capabilities**: 100/100
  - Implements a current MCP spec version (2026-07-28).

## Install

### How do I install the AI Apps That Work free tools MCP server?

AI Apps That Work free tools is a hosted endpoint at https://aiappsthatwork.com/mcp, so there is nothing to install locally. Ready-made configuration for Claude, Cursor, VS Code, Codex and 5 more is on this page, copied from each client's own documentation.

### Claude

```bash
claude mcp add --transport http com-aiappsthatwork-tools 'https://aiappsthatwork.com/mcp'
```

### Cursor

```json
{
  "mcpServers": {
    "com-aiappsthatwork-tools": {
      "url": "https://aiappsthatwork.com/mcp"
    }
  }
}
```

### VS Code

```json
{
  "servers": {
    "com-aiappsthatwork-tools": {
      "type": "http",
      "url": "https://aiappsthatwork.com/mcp"
    }
  }
}
```

### Codex

```toml
[mcp_servers.com-aiappsthatwork-tools]
url = "https://aiappsthatwork.com/mcp"
```

### opencode

```json
{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "com-aiappsthatwork-tools": {
      "type": "remote",
      "url": "https://aiappsthatwork.com/mcp",
      "enabled": true
    }
  }
}
```

### OpenClaw

```bash
openclaw mcp add com-aiappsthatwork-tools --url 'https://aiappsthatwork.com/mcp' --transport streamable-http
```

### Hermes

```yaml
mcp_servers:
  com-aiappsthatwork-tools:
    url: "https://aiappsthatwork.com/mcp"
```

### Netclaw

```json
{
  "McpServers": {
    "com-aiappsthatwork-tools": {
      "Transport": "http",
      "Url": "https://aiappsthatwork.com/mcp"
    }
  }
}
```

### Vellum

```bash
assistant mcp add com-aiappsthatwork-tools -t streamable-http -u 'https://aiappsthatwork.com/mcp'
```

### Other

```json
{
  "mcpServers": {
    "com-aiappsthatwork-tools": {
      "type": "http",
      "url": "https://aiappsthatwork.com/mcp"
    }
  }
}
```

The mcpServers block is a cross-client convention. Remote transports vary, so check your client's docs.

## Changelog

Every change recorded for this component, newest first. Days that predate change tracking, or that we cannot explain, say so: "we were watching and nothing happened" and "we were not watching" are different claims.

### 2026-11-20 (score 75, +1)

No change was recorded against any check on this day. Stability & Change Management went from 23 to 27. That category is still filling its 30-day observation window: 7 days of observed history at the previous scan, 8 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-11-18 (score 74, +1)

No change was recorded against any check on this day. Stability & Change Management went from 17 to 20. That category is still filling its 30-day observation window: 5 days of observed history at the previous scan, 6 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-11-16 (score 73, +1)

- [security] The server rewrote its instructions, which are the text every model session reads

### 2026-11-14 (score 72, +1)

- [security] The server rewrote its instructions, which are the text every model session reads

### 2026-11-13 (score 71, 0)

- [functional improvement] Stability: unverified → 0.03

### 2026-11-12 (score 71)

First indexed and scored.

## MCP tools (8)

### `real_cost_per_result` (~227 tokens)

Real cost per AI result

Prices the AI steps behind one request from current list prices (LLM tokens, images, video seconds or a fixed amount) and returns the cost per request, the real cost per result that worked first time (cost per request divided by the worked-first-time share), the monthly spend wasted on results that didn't work, and credit-pricing numbers: margin per result and the credits per request needed to break even or reach a target margin.

Input parameters:

- `pricing` (object): Credit pricing. Defaults: $10 for 1,000 credits, 1 credit per request, a 2.9% + $0.30 fee and a 70% target margin.
- `requestsPerMonth` (number, required): Requests per month, try-agains included
- `steps` (array, required): The AI steps behind one request, priced from /models.json. Their costs add up to the cost per request.
- `workedFirstTimePct` (number, required): Share of results that worked first time: no try again, undo, fix, re-ask or overrule. 1 to 100.

Output parameters:

- `costPerRequestUsd` (number): Sum of the steps' costs for one request
- `credits` (object)
- `monthlySpendUsd` (number)
- `permalink` (string): A page on this site that shows this result, for a person to open
- `pricesChecked` (string): Date the model prices were last checked
- `realCostPerResultUsd` (number): Cost per request divided by the worked-first-time share: what each result that worked costs, counting failed tries
- `requestsPerResult` (number): Requests paid for per result that worked
- `steps` (array)
- `summary` (string)
- `wastedSpendPerMonthUsd` (number): Monthly spend on results that didn't work first time

### `model_switch_cost` (~173 tokens)

Model switch cost

Compares two priced LLMs on six common task shapes (support answer, invoice extraction, lead scoring, email draft, translation, search re-ranking). Returns list prices, cost per 1,000 requests on each model, the cost ratio, the break-even (how often the new model's results must work first time, relative to the current one, to cost no more per result that works) and the current model's announced shutdown.

Input parameters:

- `from` (string, required): The current model: an LLM id from /models.json (openai/gpt-4o), an OpenRouter id or the model's name
- `to` (string, required): The new model: an LLM id from /models.json (openai/gpt-6-luna), an OpenRouter id or the model's name

Output parameters:

- `costChange` (string): The change in cost per request across the tasks
- `from` (object)
- `fromShutdown`: The current model's announced API shutdown, if any
- `note` (string|null)
- `permalink` (string): A page on this site that shows this result, for a person to open
- `pricesChecked` (string)
- `summary` (string)
- `switchTestUrl` (string|null): The browser test that runs your own prompt on both models with your own API key
- `tasks` (array)
- `to` (object)
- `tokenFactor` (number): Tokens the new model counts for text that is 1 token on the current one, where the provider publishes it; otherwise 1

### `model_prices` (~160 tokens)

AI model prices

Looks up list prices for LLMs (USD per million input, output and cached input tokens, and context size), image models (USD per image by quality and size) and video models (USD per second by resolution), with each price's source page, the date prices were checked and any announced API shutdown.

Input parameters:

- `kind` (string): llm for text models priced per token, image per image, video per second
- `limit` (integer): Most models to return
- `query` (string): A model id such as 'gpt-4o' (only exact id matches, if any), or words that must all appear in the id, name or provider, such as 'claude sonnet'. Omitted: every model.

Output parameters:

- `models` (array)
- `permalink` (string): A page on this site that shows this result, for a person to open
- `pricesChecked` (string): Date of the last full price check. Each model also has its own priceChecked and priceChanged.
- `totalMatches` (integer)

### `model_shutdowns` (~152 tokens)

AI model shutdown dates

Lists announced API shutdown dates for AI models from OpenAI, Anthropic, Google and other providers, with days left, the suggested replacement, the source page, and links to a cost comparison and a side-by-side test of the replacement where this site has them.

Input parameters:

- `includePast` (boolean): Include shutdowns whose date has passed
- `query` (string): A model id such as 'gpt-4o' (only exact id matches, if any), or words that must all appear in the model ids or provider, such as 'anthropic'. Omitted: every model.
- `today` (string): Date to count days from, YYYY-MM-DD. Omitted: today in UTC.

Output parameters:

- `count` (integer)
- `listChecked` (string): Date the list was last checked against the providers' deprecation pages
- `permalink` (string): A page on this site that shows this result, for a person to open
- `shutdowns` (array)
- `today` (string)

### `ab_test_sample_size` (~184 tokens)

Prompt or model A/B test sample size

Returns how many results each version of a prompt or model needs before a difference in how often results work first time is statistically real (two-proportion z-test, two-sided), the total across both versions, and how many days that takes at a given daily volume.

Input parameters:

- `baselinePct` (number, required): Share of version A's (the current prompt or model) results that work first time, percent (above 0, below 100)
- `confidencePct` (number): Confidence level, percent. Two-sided.
- `powerPct` (number): Chance of detecting a real difference of this size, percent
- `resultsPerDay` (number): Results per day across both versions, split evenly. Gives the number of days.
- `targetPct` (number, required): Share of version B's results that work first time, percent (above 0, below 100)

Output parameters:

- `alpha` (number): False-positive rate, 1 minus the confidence level
- `baselinePct` (number)
- `days`: Null when resultsPerDay isn't given
- `method` (string)
- `otherLifts` (array): The same test for B at +2, +5, +10 and +15 points; null where B would pass 100%
- `perVersion` (integer)
- `permalink` (string): A page on this site that shows this result, for a person to open
- `power` (number)
- `summary` (string)
- `targetPct` (number)
- `total` (integer)

### `judge_agreement` (~222 tokens)

LLM judge agreement

Compares a person's verdicts with an AI judge's verdicts on the same results. Returns the agreement rate, Cohen's kappa with confidence intervals, precision and recall per verdict, the agreement of a judge that always gives the most common verdict, a trust verdict (kappa 0.6 or more with at least 80% recall on every verdict; 0.4 to 0.6 is borderline; under 30 results is too few) and the rows where the two disagree.

Input parameters:

- `csv` (string): The same data as CSV or TSV text instead of rows: columns named human and judge (and optionally note), or the first two columns in that order.
- `rows` (array): One row per result. Common words are mapped onto two verdicts: worked (worked, pass, yes, true, 1, good, kept, accepted, ok) and didn't work (didn't work, fail, no, false, 0, bad, redo, retry, reject…

Output parameters:

- `agreement` (number): Share of results where the judge and the person gave the same verdict
- `agreementCi` (array): 95% confidence interval
- `alwaysSameVerdict`: The agreement a judge would get by always giving the most common human verdict
- `disagreementCount` (integer)
- `disagreements` (array): The first 25 results where the verdicts differ
- `kappa` (number|null): Cohen's kappa: agreement beyond chance. Null when it is undefined.
- `kappaBand` (string): Landis and Koch label for kappa
- `kappaCi`
- `matrix` (object)
- `n` (integer): Results compared
- `perVerdict` (array)
- `permalink` (string): A page on this site that shows this result, for a person to open
- `skippedRows` (integer): CSV lines without both verdicts
- `summary` (string)
- `verdict` (object)
- `warnings` (array): Problems with the labelled sample itself

### `build_judge_prompt` (~431 tokens)

Judge prompt builder

Builds a pass/fail LLM-as-a-judge prompt for one kind of AI step (make, answer, transform, extract, judge, find, predict, plan, act or learn). Returns the prompt with {{input}} and {{result}} placeholders, numbered pass conditions (the block's starter conditions unless you send your own), up to 3 pass and 3 fail examples interleaved, and a JSON or XML answer format.

Input parameters:

- `block` (string, required): What the AI step does. make: creates new content from a request (an image, a post, slides, advice); answer: replies to a question that has a right answer (a support reply, a policy question, a how-to…
- `criteria` (array): Pass conditions, one per item. Omitted: the starter conditions for the block.
- `failedExamples` (array): Real results that didn't work; labelled fail in the prompt
- `featureDescription` (string): Who uses the feature, what they ask for and what a good result gives them
- `includeReasoning` (boolean): Whether the judge writes a short reason before its verdict
- `outputFormat` (string): How the judge answers
- `workedExamples` (array): Real results that worked first time; labelled pass in the prompt

Output parameters:

- `block` (string)
- `criteria` (array): The pass conditions used
- `leftOutOfPermalink` (array): Inputs too long for the permalink, so its page shows the prompt without them
- `permalink` (string): A page on this site that shows this result, for a person to open
- `placeholders` (array)
- `prompt` (string): The judge prompt. Fill {{input}} and {{result}} for each result you check.
- `summary` (string)
- `unfilled` (object): Parts of the prompt still shown as [bracketed] placeholders

### `launch_checklist` (~233 tokens)

AI launch checklist

The 20 AI-specific launch checks for an AI app (spending caps, rate limits, server-side keys, credit charging, version tags, records of which results worked first time, real cost per result, switching models safely). Returns the checks not yet done, each with why it matters and a request for a coding agent, plus one combined request.

Input parameters:

- `done` (array): Ids of checks already done. Ids: spend-cap, rate-limits, server-keys, capped-retries, provider-budget, every-call-charged, deduct-server-side, refund-failures, one-charge-per-request, output-metadata…
- `group` (string): Only checks in one group: cost (cost and abuse), charging (charging for ai use), traceability (a record of every ai result), signals (what worked, for whom, at what cost), change (switching safely)

Output parameters:

- `combinedPrompt` (string): One message covering every remaining check, numbered; empty when none remain
- `doneCount` (integer)
- `groups` (array)
- `permalink` (string): A page on this site that shows this result, for a person to open
- `remaining` (array)
- `summary` (string)
- `total` (integer)
- `unknownIds` (array): Ids in done that aren't checks

## Diagnostics

Captured diagnostic sections: TLS, DNSSEC, Authorisation, Transports. The full working is on the page: https://verifymcp.io/servers/com-aiappsthatwork-tools/aiappsthatwork#diagnostics

## Score history

- 2026-11-21: 75
- 2026-11-20: 75
- 2026-11-19: 74
- 2026-11-18: 74
- 2026-11-17: 73
- 2026-11-16: 73
- 2026-11-15: 72
- 2026-11-14: 72
- 2026-11-13: 71
- 2026-11-12: 71

## Common questions

### What is the AI Apps That Work free tools MCP server?

AI Apps That Work free tools is an MCP server listed in the public MCP registry as com.aiappsthatwork/tools. Real cost per AI result, LLM prices, switch costs, shutdown dates, A/B test sizes and judge checks. This page covers its hosted endpoint (https://aiappsthatwork.com/mcp).

### Is the AI Apps That Work free tools MCP server safe to use?

AI Apps That Work free tools scores 75 out of 100 on VerifyMCP. That is a record of what we were able to check automatically, not an endorsement. The category breakdown on this page shows every signal behind the number, including the ones we could not confirm.

### What tools does the AI Apps That Work free tools MCP server expose?

AI Apps That Work free tools exposes 8 tools: real_cost_per_result, model_switch_cost, model_prices, model_shutdowns, ab_test_sample_size, and 3 more. Their descriptions and schemas cost roughly 1,782 tokens of context every time the server is loaded.

### Does the AI Apps That Work free tools MCP server require authentication?

No. We connected to AI Apps That Work free tools without credentials and it answered, so anything it exposes is reachable by anyone who knows the address.

### Is the AI Apps That Work free tools MCP server still maintained?

AI Apps That Work free tools is still listed as active in the MCP registry. We last reached this channel on 21 November 2026. Those dates come from our own scans of the registry and the channel itself, not from anything the publisher announced.

### What did VerifyMCP test on the AI Apps That Work free tools MCP server?

VerifyMCP ran 13 checks on AI Apps That Work free tools, covering TLS certificate, Authorization, HTTPS, HSTS header and 9 more. Supply Chain Security and Provenance & Transparency do not apply to a hosted endpoint, so they were not tested. The score does not measure everything, and the scoring documentation lists what it leaves out.

## Links

- Remote endpoint: https://aiappsthatwork.com/mcp
- Repository: https://github.com/amit-timalsina/aiappsthatwork-mcp
- Website: https://aiappsthatwork.com/connect
- Changelog RSS feed: https://verifymcp.io/servers/com-aiappsthatwork-tools/aiappsthatwork.xml
- Changelog JSON feed: https://verifymcp.io/servers/com-aiappsthatwork-tools/aiappsthatwork/changelog.json
- Trust report (JSON): https://verifymcp.io/servers/com-aiappsthatwork-tools/aiappsthatwork.json
- HTML version of this page: https://verifymcp.io/servers/com-aiappsthatwork-tools/aiappsthatwork
