# AI use cases for Think Watch Lite (ThinkWatchProject/ThinkWatch-Lite)

Repo: [ThinkWatchProject/ThinkWatch-Lite](https://github.com/ThinkWatchProject/ThinkWatch-Lite) | Use cases: 4 | Page: https://dedupedai.com/use-cases/thinkwatchproject-thinkwatch-lite/

ThinkWatch Lite is a local gateway on macOS, Windows and Linux that sits between AI clients such as Claude Code and Codex and their upstreams, recording requests and filtering keys and tool calls.

This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL.

- Connect Claude Code once, then switch upstreams in the gateway: https://dedupedai.com/use-cases/thinkwatchproject-thinkwatch-lite/connect-claude-code-once-switch-upstreams/
- Hide API keys from a relay before requests leave your machine: https://dedupedai.com/use-cases/thinkwatchproject-thinkwatch-lite/hide-api-keys-from-relay-before-sending/
- Cut off dangerous tool calls from a relay before the client runs them: https://dedupedai.com/use-cases/thinkwatchproject-thinkwatch-lite/block-dangerous-tool-calls-from-relay-answers/
- Trace each request's cost and route, then replay it on another upstream: https://dedupedai.com/use-cases/thinkwatchproject-thinkwatch-lite/compare-request-cost-and-route-across-upstreams/

## How to connect Claude Code once, then switch upstreams in the gateway with Think Watch Lite

Repo: [ThinkWatchProject/ThinkWatch-Lite](https://github.com/ThinkWatchProject/ThinkWatch-Lite) | Difficulty: Beginner | Time: 30 minutes | Basis: From the README

Point Claude Code or Codex at a local gateway once, then change providers without editing the client again.

**Who it is for:** A developer who uses Claude Code or Codex and wants to change API keys, relays or models without reconfiguring each client every time.

**The problem:** Every upstream change normally means editing the configuration of each AI client by hand, and a mistake can break the client.

**The outcome:** The client is connected to the local gateway a single time, and later upstream and model changes happen only inside the gateway.

### Steps

1. Install the app on macOS with `brew install --cask thinkwatchproject/tap/thinkwatch-lite`, or download the installer for your platform from the latest GitHub release.
2. On first launch, follow the extra step in the install guide, since the app is not signed by Apple or Microsoft.
3. Open the app and connect Claude Code or Codex in one step, reviewing the previewed change; the original file is backed up and can be restored.
4. Add your upstreams in the gateway, such as an API key, a relay or a local model.
5. Change the upstream or model in the gateway and send a test prompt from the client to confirm it answers.

### Prompt to paste into your agent

```text
Using the repo ThinkWatchProject/ThinkWatch-Lite, connect <client name, e.g. Claude Code> to the local gateway, add the upstream <upstream name> in the gateway, and switch to it without editing the client's configuration again. Expected output: a test request from the client that completes through the new upstream. Verify that the client's original config file exists as a backup.
```

### Caveat

The README says the app is not signed by Apple or Microsoft, so the first launch needs an extra step described in the install guide.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/thinkwatchproject-thinkwatch-lite/connect-claude-code-once-switch-upstreams/

## How to hide API keys from a relay before requests leave your machine with Think Watch Lite

Repo: [ThinkWatchProject/ThinkWatch-Lite](https://github.com/ThinkWatchProject/ThinkWatch-Lite) | Difficulty: Intermediate | Time: 45 minutes | Basis: From the README

Replace secrets in requests with placeholders so a relay never receives the real key or private data.

**Who it is for:** A user who sends AI requests through a third-party relay and does not want that relay to see API keys, private keys or personal identifiers.

**The problem:** A relay sees every request in full, so any key or sensitive value inside a prompt is exposed to it.

**The outcome:** API keys, private keys, JWTs, connection-string passwords and similar values are swapped for placeholders before the request leaves the machine.

### Steps

1. Install the app and connect your client through the gateway as described in the install section of the README.
2. Add the relay as an upstream in the gateway.
3. Open the secrets and privacy protection and switch it on, starting from Monitor mode to see what it would change.
4. Send a test request that contains a dummy key and check the request log to confirm the placeholder was used.
5. Switch the protection from Monitor to active once you are satisfied with the results.

### Prompt to paste into your agent

```text
Using the repo ThinkWatchProject/ThinkWatch-Lite, enable the secrets and privacy protection and route a test request containing the dummy value <dummy key> through the relay <relay name>. Expected output: the relay-side request shows a placeholder instead of the dummy key. Verify this in the request history of the gateway.
```

### Caveat

The README says each protection starts in Monitor mode and is switched on one at a time, so it does not block anything until you turn it on.

Risk, credentials: Test with dummy values first and keep the gateway's request history private, since it stores request and answer text that may include keys.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/thinkwatchproject-thinkwatch-lite/hide-api-keys-from-relay-before-sending/

## How to cut off dangerous tool calls from a relay before the client runs them with Think Watch Lite

Repo: [ThinkWatchProject/ThinkWatch-Lite](https://github.com/ThinkWatchProject/ThinkWatch-Lite) | Difficulty: Intermediate | Time: 45 minutes | Basis: From the README

Stop answers that download code, read credentials or add startup items before Claude Code or Codex executes them.

**Who it is for:** A developer who uses a relay or third-party model and wants a safety net against malicious tool calls inside its answers.

**The problem:** A relay can rewrite answers, so an answer may carry a tool call that downloads and runs code or reads private keys.

**The outcome:** Answers carrying the listed dangerous tool calls are cut off before the client can run them, and each rule can be set to its own action.

### Steps

1. Install the app and connect Claude Code or Codex to the gateway.
2. Add the relay or third-party upstream you want to protect.
3. Open the Dangerous commands protection and review its rules while it is in Monitor mode, which records and notifies.
4. Turn on the rules you want to enforce, one at a time, and set the action for each rule.
5. Run a harmless test prompt through the relay and check the request history for the recorded result.

### Prompt to paste into your agent

```text
Using the repo ThinkWatchProject/ThinkWatch-Lite, enable the Dangerous commands protection for the upstream <upstream name> and report which rules are in Monitor mode and which are enforcing. Expected output: a list of rules with their action. Verify that a test answer containing a flagged tool call is recorded in the request history.
```

### Caveat

The README lists which kinds of tool calls are covered, so other risky commands may pass through unless you add a rule for them.

Risk, remote script: Protection only covers the rules listed in the README, so review the rule list yourself before trusting a relay with untrusted tool use.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/thinkwatchproject-thinkwatch-lite/block-dangerous-tool-calls-from-relay-answers/

## How to trace each request's cost and route, then replay it on another upstream with Think Watch Lite

Repo: [ThinkWatchProject/ThinkWatch-Lite](https://github.com/ThinkWatchProject/ThinkWatch-Lite) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README

See which upstream served a request, what it cost, and compare the answer with a replay on a different upstream.

**Who it is for:** A user who runs the same model through several upstreams and wants to know which one is cheaper or more reliable.

**The problem:** Without a record of routes and costs, it is hard to tell which upstream answered a request or whether a price is accurate.

**The outcome:** Each request shows the rule it matched, the upstreams it tried, format conversions and its cost, and a finished request can be replayed side by side on another upstream.

### Steps

1. Install the app and route your client through the gateway.
2. Add two or more upstreams that serve the same model.
3. Send a few requests from the client and open the request history.
4. Pick a finished request and replay it against a different upstream.
5. Compare the two answers and costs side by side, noting that estimated amounts are marked.

### Prompt to paste into your agent

```text
Using the repo ThinkWatchProject/ThinkWatch-Lite, send one prompt through the upstream <upstream A> and replay it on <upstream B>, then report both costs and whether each cost is marked as estimated. Expected output: a side-by-side comparison of the two answers and costs. Verify that requests without a price are counted separately, not as zero.
```

### Caveat

The README says requests without a known price are counted separately instead of as zero, so totals may not cover every request.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/thinkwatchproject-thinkwatch-lite/compare-request-cost-and-route-across-upstreams/