### Install and Run GPTScript on Windows

Source: https://docs.gptscript.ai/

Installation method for Windows using the winget package manager.

```powershell
winget install gptscript-ai.gptscript  
gptscript github.com/gptscript-ai/llm-basics-demo  
```

--------------------------------

### Install and Run GPTScript via Homebrew Core

Source: https://docs.gptscript.ai/

Standard Homebrew installation, though note that the core formula may be outdated compared to the tap.

```bash
brew install gptscript  
gptscript github.com/gptscript-ai/llm-basics-demo  
```

--------------------------------

### Tool Body Instructions

Source: https://docs.gptscript.ai/tools/gpt-file-reference

Example showing both natural language prompts and shell command execution within a tool body.

```text
Name: echo-ai  
Description: A tool that echos the input  
Parameter: input: The input  
  
Just return only "${input}"  
  
---
Name: echo-command  
Description: A tool that echos the input  
Parameter: input: The input  
  
#!/bin/bash  
  
echo "${input}"  

```

--------------------------------

### Example output file

Source: https://docs.gptscript.ai/examples/workflow

The resulting markdown file generated by the script.

```markdown
## Jun 4, 2024 [Acorn Labs Tweet](https://x.com/acornlabs/status/1798063732394000559)  
  
Our latest release of #GPTScript v0.7 is available now. Key updates include OpenAPI v2 support, workspace tools replacement, and a chat history context tool.  
  
### References  
  
- [Acorn | Introducing GPTScript v0.7](https://buff.ly/3V1KiWw): The page is a blog post introducing GPTScript v0.7, highlighting its new features and fixes. Key updates include support for OpenAPI v2, removal and replacement of workspace tools, introduction of a chat history context tool, and several housekeeping changes like defaulting to the gpt-4o model and logging token usage. The post also mentions upcoming features in v0.8 and provides links to related articles and resources.  
  
## Jun 4, 2024 [Acorn Labs Tweet](https://x.com/acornlabs/status/1797998244447900084)  
  
Use #GPTScript with #Python to build the front end of this AI-powered YouTube generator. Part of this series walks you through it all.  
  
### References  
  
- [Tutorial on building an AI-powered YouTube title and thumbnail generator (Part 1)](https://buff.ly/4aGJYCD): Covers setting up the front end with Flask, modifying the script, and creating templates for the front page and result page.  
- [Tutorial on building an AI-powered YouTube title and thumbnail generator (Part 2)](https://buff.ly/49MSK1k): Covers the installation and setup of GPTScript, creating a basic script, handling command-line arguments, generating a YouTube title, and generating thumbnails using DALL-E 3.  
```

--------------------------------

### Install and Run GPTScript via Homebrew Tap

Source: https://docs.gptscript.ai/

Recommended method for MacOS and Linux to ensure the latest version is installed.

```bash
brew install gptscript   
gptscript github.com/gptscript-ai/llm-basics-demo  
```

--------------------------------

### Specify Workspace Directory

Source: https://docs.gptscript.ai/examples/local-files

Example command syntax for specifying a custom workspace directory.

```bash
gptscript --workspace=/Your/path/here ...
```

--------------------------------

### Define a Daemon Tool with Node.js

Source: https://docs.gptscript.ai/tools/daemon

Example of a GPTScript tool definition using #!sys.daemon to launch a Node.js HTTP server.

```text
Tools: my-daemon  
Param: first: the first parameter  
Param: second: the second parameter  
  
#!http://my-daemon.daemon.gptscript.local/myPath  
  
---  
Name: my-daemon  
  
#!sys.daemon node  
  
const http = require('http');  
  
const server = http.createServer((req, res) => {  
    if (req.method === 'GET' || req.method === 'POST') {  
        // Extract the path from the request URL  
        const path = req.url;  
  
        let body = '';  
  
        req.on('data', chunk => {  
          body += chunk.toString();  
        })  
  
        // Respond with the path and body  
        req.on('end', () => {  
          res.writeHead(200, { 'Content-Type': 'text/plain' });  
          res.write(`Body: ${body}\n`);  
          res.end(`Path: ${path}`);  
        })  
    } else {  
        res.writeHead(405, { 'Content-Type': 'text/plain' });  
        res.end('Method Not Allowed');  
    }  
});  
  
const PORT = process.env.PORT || 3000;  
server.listen(PORT, () => {  
    console.log(`Server is listening on port ${PORT}`);  
});
```

--------------------------------

### Install GPTScript via shell

Source: https://docs.gptscript.ai/faqs

Use this command on macOS or Linux if Homebrew is not available.

```bash
curl https://get.gptscript.ai/install.sh | sh
```

--------------------------------

### Tool Definition Preamble

Source: https://docs.gptscript.ai/tools/gpt-file-reference

Example of a tool definition including comments and parameter declarations.

```text
Name: tool-name  
# This is a comment in the preamble.  
Description: Tool description  
# This tool can invoke tool1 or tool2 if needed  
Tools: tool1, tool2  
Param: param1: The description of param1  
  
Tool instructions go here.  

```

--------------------------------

### GPTScript File Structure

Source: https://docs.gptscript.ai/tools/gpt-file-reference

Example of a GPTScript file containing multiple tools separated by triple dashes.

```text
Name: tool1  
Description: This is tool1  
  
Do sample tool stuff.  
  
---
Name: tool2  
Description: This is tool2  
  
Do more sample tool stuff.  

```

--------------------------------

### Referencing External Tools

Source: https://docs.gptscript.ai/tools/using

Examples of importing tools from remote URLs or local files.

```gptscript
tools: https://raw.githubusercontent.com/gptscript-ai/gptscript/refs/heads/main/examples/echo.gpt  
  
Echo the phrase "Hello, World!".
```

```gptscript
tools: echo.gpt  
  
Echo the phrase "Hello, World!".
```

--------------------------------

### Use OpenAPI Petstore as a Tool

Source: https://docs.gptscript.ai/tools/openapi

Example of referencing an OpenAPI definition URL directly within a GPTScript tool definition.

```text
Tools: https://petstore.gptscript-demos.ai/openapi  
  
List all the pets. What are the names of the dogs?  
```

--------------------------------

### GPTScript execution output

Source: https://docs.gptscript.ai/tools/how-it-works

The expected output after running the provided GPTScript example.

```text
OUTPUT:  
  
Bob said, "Thanks for asking 'How are you doing?', I'm doing great fellow friendly AI tool!"  
```

--------------------------------

### Define credential in GPTScript

Source: https://docs.gptscript.ai/tools/credential-tools

Example of referencing a credential tool within a script.

```gptscript
Credential: my-credential-tool.gpt as myCred  
  
<tool stuff here>
```

--------------------------------

### Define remote tools in GPTScript

Source: https://docs.gptscript.ai/faqs

Example syntax for including a remote tool or context within a script.

```text
Tools: github.com/gptscript-ai/browser
```

--------------------------------

### Define Security Schemes

Source: https://docs.gptscript.ai/tools/openapi

Example of defining multiple security schemes globally in an OpenAPI document.

```yaml
servers:  
  - url: https://api.example.com/v1  
components:  
  securitySchemes:  
    MyBasic:  
      type: http  
      scheme: basic  
  
    MyAPIKey:  
      type: apiKey  
      in: header  
      name: X-API-Key  
security:  
  - MyBasic: []  
  - MyAPIKey: []  
# The rest of the document defines paths, etc.  
```

--------------------------------

### Clean and format data

Source: https://docs.gptscript.ai/examples/local-files

Example interaction where the assistant reads a text file and converts it into a structured CSV format.

```text
> whats in the key contacts file?  
  
  The  key_contacts.txt  file contains the following information:  
  
    Contains location, name & email  
    Scranton - Michael michael@gptscript-fakeco.com  
    Stamford Martin marty@gptscript-fakeco.com  
    Nashua - Rachael, rach@gptscript-fakeco.com  
    Utica,Ross - ross@gptscript-fakeco.com  
    Albany    Allen    allen@gptscript-fakeco.com  
    Akron;Pam;pam@gptscript-fakeco.com  
    Camden. Kali. kali@gptscript-fakeco.com  
    Yonkers - Meredith - meredith@gptscript-fakeco.com  
    Rochester Kevin  
    Syracuse William will@gptscript-fakeco.com  
  
  Is there anything specific you would like to do with this information?  
  
> That looks bad. Can you make a cleaned up csv of it?  
  
  Sure, I'll clean up the data and create a CSV file for you. Here's the cleaned-up data in CSV format:  
  
    Location,Name,Email  
    Scranton,Michael,michael@gptscript-fakeco.com  
    Stamford,Martin,marty@gptscript-fakeco.com  
    Nashua,Rachael,rach@gptscript-fakeco.com  
    Utica,Ross,ross@gptscript-fakeco.com  
    Albany,Allen,allen@gptscript-fakeco.com  
    Akron,Pam,pam@gptscript-fakeco.com  
    Camden,Kali,kali@gptscript-fakeco.com  
    Yonkers,Meredith,meredith@gptscript-fakeco.com  
    Rochester,Kevin,  
    Syracuse,William,will@gptscript-fakeco.com  
  
  I will now create a CSV file with this cleaned-up data.
```

--------------------------------

### Example Output from Filtered Tool

Source: https://docs.gptscript.ai/tools/input-output-filters

The resulting text after the output filter has processed the poem.

```text
OUTPUT:  
  
In orchards where the sunlight gleams, Among the leaves, in golden beams, The oranges hang on branches high, A feast for both the heart and eye.  
Their skins, a palette rich and bright, In hues of red and green delight, With every bite, a crisp surprise, A taste of autumn, pure and wise.  
From pies to cider, sweet and bold, Their stories through the seasons told, In every crunch, a memory, Of nature's gift, so wild and free.  
Oh, oranges, treasures of the earth, In every form, you bring us mirth, A simple fruit, yet so profound, In you, a world of joy is found.
```

--------------------------------

### Cross-reference data files

Source: https://docs.gptscript.ai/examples/local-files

Example interaction where the assistant retrieves specific sales data from an Excel file based on a location identified in previous steps.

```text
> what were sales like for Kevin's location?  
  
  To find out sales for Kevin's location, I will need to check the  yoy_sales.xlsx  file for any relevant data. Let me take a  
  look at the contents of that file.  
...  
  The sales data for Rochester, where Kevin is located, is as follows:  
  
    LOCATION  │ PRODUCT SKU │ MARCH 2023 SALES │ MARCH 2024 SALES │ YOY SALES GROWTH (%)  
  ────────────┼─────────────┼──────────────────┼──────────────────┼───────────────────────  
    Rochester │ SKU_001     │ $803.91          │ $843.10          │ 12.12%  
    Rochester │ SKU_002     │ $3,013.36        │ $3,159.02        │ 13.13%  
    Rochester │ SKU_003     │ $1,879.24        │ $1,968.20        │ 12.15%  
    Rochester │ SKU_004     │ $4,989.81        │ $5,239.30        │ 13.88%  
    Rochester │ SKU_005     │ $4,334.03        │ $4,550.73        │ 12.70%  
    Rochester │ SKU_006     │ $2,143.76        │ $2,251.95        │ 13.07%  
    Rochester │ SKU_007     │ $7,976.54        │ $8,375.37        │ 12.23%  
    Rochester │ SKU_008     │ $313.21          │ $328.87          │ 13.80%  
    Rochester │ SKU_009     │ $671.70          │ $705.29          │ 12.91%  
    Rochester │ SKU_010     │ $4,084.94        │ $4,289.19        │ 13.03%  
  
  Is there anything else you would like to know or do with this data?
```

--------------------------------

### Identify and fix data gaps

Source: https://docs.gptscript.ai/examples/local-files

Example interaction where the assistant identifies missing information in a CSV and updates it based on user instructions.

```text
> is there any missing data in that csv?  
...  
  Yes, there is missing data in the CSV file. Specifically, the email for Kevin in Rochester is missing.  
  
  Would you like to update this information or perform any other actions?  
> yes set his email to kev@ and use the same domain as the others  
  
  I will update Kevin's email to  kev@gptscript-fakeco.com  in the CSV file.
```

--------------------------------

### Run the workflow demo

Source: https://docs.gptscript.ai/examples/workflow

Execute the pre-built workflow demo directly from the repository.

```bash
gptscript github.com/gptscript-ai/workflow-demo
```

--------------------------------

### Initialize the script file

Source: https://docs.gptscript.ai/examples/workflow

Create a new GPTScript file named workflow-demo.gpt to begin development.

```bash
vim workflow-demo.gpt
```

--------------------------------

### Define the entrypoint tool

Source: https://docs.gptscript.ai/examples/cli

Configure the main agent tool with context references and chat capabilities.

```text
Context: learn-gh
Context: github.com/gptscript-ai/context/cli
Chat: true

You have the gh cli available to you. Use it to accomplish the tasks that the user asks of you.
```

--------------------------------

### Define a context provider with parameters

Source: https://docs.gptscript.ai/tools/context

Create a context tool that accepts parameters to customize its output.

```bash
# context_with_param.gpt  
Param: tone: the tone to use when responding to the user's request  
  
#!/bin/bash  
echo "Respond to the user's request in a ${tone} tone."  
  

```

--------------------------------

### Clone Sample Data Repository

Source: https://docs.gptscript.ai/examples/local-files

Downloads the sample data repository and navigates into the directory.

```bash
git clone https://github.com/gptscript-ai/local-files-demo.git
cd local-files-demo
```

--------------------------------

### Initialize the GPTScript file

Source: https://docs.gptscript.ai/examples/cli

Create a new file to house the GPTScript configuration.

```bash
vim cli-demo.gpt
```

--------------------------------

### Define tool configuration

Source: https://docs.gptscript.ai/tools/authoring

Create a tool.gpt file to specify metadata, parameters, and the execution command.

```text
Description: Returns the contents of a webpage.  
Param: url: The URL of the webpage.  
  
#!/usr/bin/env python3 ${GPTSCRIPT_TOOL_DIR}/tool.py  
```

--------------------------------

### Go project file structure

Source: https://docs.gptscript.ai/tools/code-tool-guidelines

Recommended layout for Go tools, requiring a go.mod file.

```text
.  
├── go.mod  
├── go.sum  
├── main.go  
└── tool.gpt  
```

--------------------------------

### List models from a provider

Source: https://docs.gptscript.ai/alternative-model-providers

Use the --list-models flag with the provider repository URL to see available models.

```bash
gptscript --list-models github.com/gptscript-ai/claude3-anthropic-provider  
```

--------------------------------

### Configure Daemon Readiness Path

Source: https://docs.gptscript.ai/tools/daemon

Syntax for specifying a custom readiness check path for the daemon server.

```text
#!sys.daemon (path=/api/ready) node  
  
// (node script here)
```

--------------------------------

### Define the workflow entrypoint

Source: https://docs.gptscript.ai/examples/workflow

Configure the main tool to import dependencies, enforce sequential processing, and specify target URLs for summarization.

```gptscript
Tools: sys.write, summarize-tweet  
Description: Summarize tweets  
  
Always summarize tweets synchronously in the order they appear.  
Never summarize tweets in parallel.  
Write all the summaries for the tweets at the following URLs to `tweets.md`:  
- https://x.com/acornlabs/status/1798063732394000559  
- https://x.com/acornlabs/status/1797998244447900084
```

--------------------------------

### Using Packaged GitHub Tools

Source: https://docs.gptscript.ai/tools/using

Demonstrates referencing tools hosted on GitHub using their repository path.

```gptscript
tools: github.com/gptscript-ai/dalle-image-generation, github.com/gptscript-ai/gpt4-v-vision, sys.read  
  
Generate an image of a city skyline at night and write the resulting image to a file called city_skyline.png.  
  
Take this image and write a description of it in the style of pirate.
```

--------------------------------

### Show credential command syntax

Source: https://docs.gptscript.ai/command-line-reference/gptscript_credential_show

Basic usage syntax for displaying a stored credential.

```bash
gptscript credential show <credential name> [flags]
```

--------------------------------

### Show command options

Source: https://docs.gptscript.ai/command-line-reference/gptscript_credential_show

Available flags for the show command.

```bash
-h, --help   help for show
```

--------------------------------

### Utilizing System Tools

Source: https://docs.gptscript.ai/tools/using

Demonstrates invoking built-in system tools like sys.read and sys.write within a script.

```gptscript
tools: sys.read, sys.write  
  
Read all of the files in my current directory, do not recurse over any subdirectories, and give me a description of this directory's contents.
```

--------------------------------

### Define the learn-gh context tool

Source: https://docs.gptscript.ai/examples/cli

Provide the LLM with CLI help documentation by executing a bash script that outputs command usage information.

```bash
---
Name: learn-gh

#!/usr/bin/env bash

echo "The following is the help text for the gh cli and some of its sub-commands. Use these when figuring out how to construct new commands. Note that the --search flag is used for filtering and sorting as well; there is no dedicated --sort flag."
gh --help
gh repo --help
gh issue --help
gh issue list --help
gh issue create --help
gh issue comment --help
gh issue delete --help
gh issue edit --help
gh pr --help
gh pr create --help
gh pr checkout --help
gh release --help
gh release create --help
```

--------------------------------

### Create Assistant File

Source: https://docs.gptscript.ai/examples/local-files

Opens a new GPTScript file for editing.

```bash
vim data-assistant.gpt
```

--------------------------------

### Implement a programmatic tool in Python

Source: https://docs.gptscript.ai/tools/how-it-works

Demonstrates implementing a tool using a Python script instead of a natural language prompt.

```python
Tools: bob  
  
Ask Bob how he is doing and let me know exactly what he said.  
  
---  
Name: bob  
Description: I'm Bob, a friendly guy.  
Param: question: The question to ask Bob.  
  
#!python3  
  
import os  
  
print(f"Thanks for asking {os.environ['question']}, I'm doing great fellow friendly AI tool!")  
```

--------------------------------

### Import a shared tool

Source: https://docs.gptscript.ai/tools/authoring

Use an existing tool from a GitHub repository in your own script.

```text
tools: github.com/gptscript-ai/dalle-image-generation  
  
Generate an image of a city skyline at night.  
```

--------------------------------

### List models from an OpenAI-compatible API

Source: https://docs.gptscript.ai/alternative-model-providers

Use the --list-models flag with the API base URL to list available models.

```bash
gptscript --list-models https://api.mistral.ai/v1  
```

--------------------------------

### Define static context text

Source: https://docs.gptscript.ai/tools/context

Create a text file to serve as a static context provider.

```text
# my-context.txt  
  
You have access to run commands on the user's system. Please ask for confirmation from the user before running a command.  
  

```

--------------------------------

### Python project file structure

Source: https://docs.gptscript.ai/tools/code-tool-guidelines

Recommended layout for Python tools with dependencies.

```text
.  
├── requirements.txt  
├── tool.py  
└── tool.gpt  
```

--------------------------------

### Node.js project file structure

Source: https://docs.gptscript.ai/tools/code-tool-guidelines

Recommended layout for Node.js tools with dependencies.

```text
.  
├── package.json  
├── tool.js  
└── tool.gpt  
```

--------------------------------

### Use the generic credential tool

Source: https://docs.gptscript.ai/tools/credential-tools

Utilize the built-in generic credential tool to prompt for a single environment variable.

```text
Credentials: github.com/gptscript-ai/credential as myCredentialName with MY_ENV_VAR as env and "this message will be displayed to the user" as message and key as field  
  
(tool stuff here)
```

--------------------------------

### Create a credential provider tool

Source: https://docs.gptscript.ai/tools/credential-tools

A bash-based credential provider that prompts the user for input and outputs the result in the required JSON format.

```bash
# my-credential-tool.gpt  
Name: my-credential-tool  
  
#!/usr/bin/env bash  
  
output=$(gptscript -q --disable-cache sys.prompt '{"message":"Please enter your fake credential.","fields":"credential","sensitive":"true"}')  
credential=$(echo $output | jq -r '.credential')  
echo "{\"env\":{\"MY_ENV_VAR\":\"$credential\"}}"
```

--------------------------------

### Command Options

Source: https://docs.gptscript.ai/command-line-reference/gptscript_getenv

Available flags for the getenv command.

```bash
  -h, --help   help for getenv
```

--------------------------------

### Usage Syntax

Source: https://docs.gptscript.ai/command-line-reference/gptscript_getenv

Basic command structure for retrieving an environment variable.

```bash
gptscript getenv [flags] KEY [DEFAULT]
```

--------------------------------

### Command options

Source: https://docs.gptscript.ai/command-line-reference/gptscript_fmt

Available flags for the fmt command.

```bash
  -h, --help    help for fmt  
  -w, --write   Write output to file instead of stdout ($GPTSCRIPT_FMT_WRITE)
```

--------------------------------

### Run Local Files Demo

Source: https://docs.gptscript.ai/examples/local-files

Executes the local-files-demo script with the specified workspace directory.

```bash
gptscript --workspace=~/Documents github.com/gptscript-ai/local-files-demo
```

--------------------------------

### Define tool dependencies

Source: https://docs.gptscript.ai/tools/authoring

List required Python packages in a requirements.txt file.

```text
requests  
```

--------------------------------

### List credentials with stacked contexts

Source: https://docs.gptscript.ai/tools/credential-tools

Displays credentials from multiple contexts, showing only the highest priority version if names collide.

```bash
gptscript --credential-context one,two credentials
```

--------------------------------

### Basic gptscript fmt usage

Source: https://docs.gptscript.ai/command-line-reference/gptscript_fmt

The standard syntax for invoking the format command.

```bash
gptscript fmt [flags]
```

--------------------------------

### Run script with stacked credential contexts

Source: https://docs.gptscript.ai/tools/credential-tools

Specifies two contexts where the first context has higher priority for credential resolution and storage.

```bash
gptscript --credential-context one,two my-script.gpt
```

--------------------------------

### Execute a tool from the command line

Source: https://docs.gptscript.ai/tools/authoring

Run a tool directly using the gptscript CLI with parameters provided as a JSON string.

```bash
gptscript github.com/<user>/<repo name> '{"url": "https://github.com"}'  
```

--------------------------------

### Define a context provider tool with exported tools

Source: https://docs.gptscript.ai/tools/context

Create a GPTScript tool that exports tools and provides dynamic context output.

```gptscript
# my-context-tool.gpt  
Share Tools: sys.exec  
  
#!sys.echo  
You have access to run commands on the user's system. Please ask for confirmation from the user before running a command.  
  

```

--------------------------------

### Execute script with parameters

Source: https://docs.gptscript.ai/tools/context

Run a GPTScript file and provide parameter values via JSON.

```bash
gptscript tool.gpt '{"tone": "obnoxious"}'  

```

--------------------------------

### Run GPTScript with a workspace

Source: https://docs.gptscript.ai/examples/local-files

Execute the assistant by specifying a local directory as the workspace.

```bash
gptscript --workspace=./sample-data data-assistant.gpt
```

--------------------------------

### Use a context provider tool

Source: https://docs.gptscript.ai/tools/context

Reference a GPTScript context tool in a script.

```gptscript
Context: ./my-context-tool.gpt  
Tools: sys.write  
  
Which processes on my system are using the most memory? Write their PIDs to a file called pids.txt.  

```

--------------------------------

### Inherited parent command options

Source: https://docs.gptscript.ai/command-line-reference/gptscript_fmt

Global configuration options inherited from the parent gptscript command.

```bash
      --cache-dir string                Directory to store cache (default: $XDG_CACHE_HOME/gptscript) ($GPTSCRIPT_CACHE_DIR)  
  -C, --chdir string                    Change current working directory ($GPTSCRIPT_CHDIR)  
      --color                           Use color in output (default true) ($GPTSCRIPT_COLOR)  
      --config string                   Path to GPTScript config file ($GPTSCRIPT_CONFIG)  
      --confirm                         Prompt before running potentially dangerous commands ($GPTSCRIPT_CONFIRM)  
      --credential-context strings      Context name(s) in which to store credentials ($GPTSCRIPT_CREDENTIAL_CONTEXT)  
      --credential-override strings     Credentials to override (ex: --credential-override github.com/example/cred-tool:API_TOKEN=1234) ($GPTSCRIPT_CREDENTIAL_OVERRIDE)  
      --debug                           Enable debug logging ($GPTSCRIPT_DEBUG)  
      --debug-messages                  Enable logging of chat completion calls ($GPTSCRIPT_DEBUG_MESSAGES)  
      --default-model string            Default LLM model to use ($GPTSCRIPT_DEFAULT_MODEL) (default "gpt-4o")  
      --default-model-provider string   Default LLM model provider to use, this will override OpenAI settings ($GPTSCRIPT_DEFAULT_MODEL_PROVIDER)  
      --disable-cache                   Disable caching of LLM API responses ($GPTSCRIPT_DISABLE_CACHE)  
      --dump-state string               Dump the internal execution state to a file ($GPTSCRIPT_DUMP_STATE)  
      --events-stream-to string         Stream events to this location, could be a file descriptor/handle (e.g. fd://2), filename, or named pipe (e.g. \\.\pipe\my-pipe) ($GPTSCRIPT_EVENTS_STREAM_TO)  
  -f, --input string                    Read input from a file ("-" for stdin) ($GPTSCRIPT_INPUT_FILE)  
      --no-trunc                        Do not truncate long log messages ($GPTSCRIPT_NO_TRUNC)  
      --openai-api-key string           OpenAI API KEY ($OPENAI_API_KEY)  
      --openai-base-url string          OpenAI base URL ($OPENAI_BASE_URL)  
      --openai-org-id string            OpenAI organization ID ($OPENAI_ORG_ID)  
  -o, --output string                   Save output to a file, or - for stdout ($GPTSCRIPT_OUTPUT)  
  -q, --quiet                           No output logging (set --quiet=false to force on even when there is no TTY) ($GPTSCRIPT_QUIET)  
      --system-tools-dir string         Directory that contains system managed tool for which GPTScript will not manage the runtime ($GPTSCRIPT_SYSTEM_TOOLS_DIR)  
      --workspace string                Directory to use for the workspace, if specified it will not be deleted on exit ($GPTSCRIPT_WORKSPACE)
```

--------------------------------

### Execute the chatbot script

Source: https://docs.gptscript.ai/examples/api

Run the configured GPTScript file to initiate the interactive chat session.

```bash
gptscript tool.gpt
```

--------------------------------

### List stored credentials command

Source: https://docs.gptscript.ai/command-line-reference/gptscript_credential

Basic syntax for listing credentials using the gptscript CLI.

```bash
gptscript credential [flags]
```

--------------------------------

### Define DigitalOcean Bot tool

Source: https://docs.gptscript.ai/examples/api

Create a tool.gpt file to configure the chatbot with specific API tools and behavioral instructions.

```text
Name: DigitalOcean Bot  
Chat: true  
Tools: droplets* from openapi.yaml  
Tools: databases* from openapi.yaml  
Tools: images* from openapi.yaml  
Tools: regions_list from openapi.yaml  
Tools: tags* from openapi.yaml  
Tools: sizes_list from openapi.yaml  
Tools: sshKeys_list from openapi.yaml  
Tools: sys.time.now  
  
You are a helpful assistant with access to the DigitalOcean API to manage droplets and databases.  
Before creating, updating, or deleting anything, tell the user about the exact action you are going to take, and get their confirmation.  
Start the conversation by asking the user how you can help.
```

--------------------------------

### Create a Python tool script

Source: https://docs.gptscript.ai/tools/authoring

Define the logic for a tool in a Python file that reads environment variables.

```python
import os  
import requests  
  
print(requests.get(os.getenv("url")).text)  
```

--------------------------------

### Define Credential Tool Parameters

Source: https://docs.gptscript.ai/tools/credential-tools

Defines a credential tool with specific parameters and a bash script to output the resulting environment variables.

```bash
Name: my-credential-tool  
Parameter: env: the environment variable to set  
Parameter: val: the value to set it to  
  
#!/usr/bin/env bash  
  
echo "{\"env\":{\"$ENV\":\"$VAL\"}}"
```

--------------------------------

### Execute the GPTScript

Source: https://docs.gptscript.ai/examples/cli

Run the defined script using the gptscript CLI.

```bash
gptscript cli-demo.gpt
```

--------------------------------

### Specify multiple credential tools

Source: https://docs.gptscript.ai/tools/credential-tools

Use multiple Credentials directives on separate lines to load multiple credential providers.

```text
Credentials: credential-tool-1.gpt  
Credentials: credential-tool-2.gpt  
  
(tool stuff here)
```

--------------------------------

### Define GPTScript tools

Source: https://docs.gptscript.ai/tools/how-it-works

Defines an entrypoint tool that calls a secondary tool named 'bob' using natural language prompts.

```gptscript
Tools: bob  
  
Ask Bob how he is doing and let me know exactly what he said.  
  
---  
Name: bob  
Description: I'm Bob, a friendly guy.  
Param: question: The question to ask Bob.  
  
When asked how I am doing, respond with "Thanks for asking "${question}", I'm doing great fellow friendly AI tool!"  
```

--------------------------------

### Call an external Go program

Source: https://docs.gptscript.ai/tools/code-tool-guidelines

Go tools must use the specific path format shown for the tool body.

```text
Name: my-go-tool  
  
#!${GPTSCRIPT_TOOL_DIR}/bin/gptscript-go-tool  
```

--------------------------------

### Reference Credential Tool with Parameters

Source: https://docs.gptscript.ai/tools/credential-tools

Sets specific parameter values when referencing a credential tool in a script.

```text
Credential: my-credential-tool.gpt with MY_ENV_VAR as env and "my value" as val  
  
(tool stuff here)
```

--------------------------------

### Show specific credential with stacked contexts

Source: https://docs.gptscript.ai/tools/credential-tools

Searches for a specific credential across multiple contexts based on priority.

```bash
gptscript --credential-context one,two credential show myCred
```

--------------------------------

### Use a credential provider tool in a script

Source: https://docs.gptscript.ai/tools/credential-tools

Reference a credential provider tool using the Credentials directive to make environment variables available to the script.

```bash
Credentials: my-credential-tool.gpt as myCred  
  
#!/usr/bin/env bash  
  
echo "The value of MY_ENV_VAR is $MY_ENV_VAR"
```

--------------------------------

### Credential command options

Source: https://docs.gptscript.ai/command-line-reference/gptscript_credential

Available flags for the credential command to control output and context scope.

```bash
      --all-contexts    List credentials for all contexts ($GPTSCRIPT_CREDENTIAL_ALL_CONTEXTS)  
  -h, --help            help for credential  
      --show-env-vars   Show names of environment variables in each credential ($GPTSCRIPT_CREDENTIAL_SHOW_ENV_VARS)
```

--------------------------------

### Reference workspace context tool

Source: https://docs.gptscript.ai/tools/workspace

Adds workspace awareness to LLM-based tools by referencing the workspace context.

```text
Context: github.com/gptscript-ai/context/workspace
```

--------------------------------

### Require Multiple Security Schemes

Source: https://docs.gptscript.ai/tools/openapi

Shows how to define security requirements as a logical AND by grouping them in the same array entry.

```yaml
security:  
  - MyBasic: []  
    MyAPIKey: []  
```

--------------------------------

### Authenticate with GitHub using the gh CLI

Source: https://docs.gptscript.ai/faqs

Set the GITHUB_AUTH_TOKEN environment variable using an existing GitHub CLI token to avoid rate limits.

```bash
export GITHUB_AUTH_TOKEN="$(gh auth token)"  
```

--------------------------------

### GPTScript Execution Syntax

Source: https://docs.gptscript.ai/command-line-reference/gptscript

The basic command structure for executing a GPTScript program file.

```bash
gptscript [flags] PROGRAM_FILE [INPUT...]
```

--------------------------------

### Download DigitalOcean OpenAPI definition

Source: https://docs.gptscript.ai/examples/api

Use curl to fetch the OpenAPI YAML file locally for easier reference in GPTScript.

```bash
curl -o openapi.yaml -L https://api-engineering.nyc3.cdn.digitaloceanspaces.com/spec-ci/DigitalOcean-public.v2.yaml
```

--------------------------------

### Defining In-Script Tools

Source: https://docs.gptscript.ai/tools/using

Shows how to define a custom tool directly within the main script file for modularity.

```gptscript
tools: random-number  
  
Select a number at random and, based on the results, write a poem about it.  
  
---  
name: random-number  
description: Generate a random number between 1 and 100.  
  
Select a number at random between 1 and 100 and return only the number.
```

--------------------------------

### Execute GPTScript with Daemon

Source: https://docs.gptscript.ai/tools/daemon

Command to run the script and the expected output from the daemon tool.

```bash
gptscript script.gpt '{"first":"hello","second":"world"}'
```

```text
OUTPUT:  
  
Body: {"first":"hello","second":"world"}  
Path: /myPath
```

--------------------------------

### Credential override via key-value pairs

Source: https://docs.gptscript.ai/tools/credential-tools

Sets specific environment variables for a credential tool using explicit values.

```text
toolA:ENV_VAR_1=value1,ENV_VAR_2=value2
```

--------------------------------

### Invoke a tool via GPTScript

Source: https://docs.gptscript.ai/tools/authoring

Reference a GitHub-hosted tool within a GPTScript file.

```text
Tools: github.com/<user>/<repo name>  
  
Get the contents of https://github.com  
```

--------------------------------

### Inherited credential options

Source: https://docs.gptscript.ai/command-line-reference/gptscript_credential

Global flags inherited from parent commands for credential context management.

```bash
      --credential-context strings   Context name(s) in which to store credentials ($GPTSCRIPT_CREDENTIAL_CONTEXT)
```

--------------------------------

### Reference an OpenAI-compatible API model

Source: https://docs.gptscript.ai/alternative-model-providers

Specify the model name and the base URL of the OpenAI-compatible API.

```text
model: mistral-large-latest from https://api.mistral.ai/v1  
  
Say hello world  
```

--------------------------------

### Execute a GPTScript file

Source: https://docs.gptscript.ai/examples/workflow

Command to run the defined workflow script.

```bash
gpscript workflow-demo.gpt  
```

--------------------------------

### Use context with parameters

Source: https://docs.gptscript.ai/tools/context

Pass parameters to a context tool within a script.

```gptscript
# tool.gpt  
Context: ./context_with_param.gpt with ${tone} as tone  
Param: tone: the tone to use when responding to the user's request  
Tools: sys.http.html2text  
  
What are the top stories on Hacker News right now?  
  

```

--------------------------------

### Define an inline Node.js tool

Source: https://docs.gptscript.ai/tools/code-tool-guidelines

Use for simple Node.js scripts without external dependencies.

```text
Name: my-node-tool  
  
#!node  
  
console.log('hello world')  
```

--------------------------------

### GPTScript Configuration Flags

Source: https://docs.gptscript.ai/command-line-reference/gptscript

A comprehensive list of flags used to configure the GPTScript runtime environment, including cache settings, model selection, and debugging options.

```bash
      --cache-dir string                    Directory to store cache (default: $XDG_CACHE_HOME/gptscript) ($GPTSCRIPT_CACHE_DIR)  
      --chat-state string                   The chat state to continue, or null to start a new chat and return the state ($GPTSCRIPT_CHAT_STATE)  
  -C, --chdir string                        Change current working directory ($GPTSCRIPT_CHDIR)  
      --color                               Use color in output (default true) ($GPTSCRIPT_COLOR)  
      --config string                       Path to GPTScript config file ($GPTSCRIPT_CONFIG)  
      --confirm                             Prompt before running potentially dangerous commands ($GPTSCRIPT_CONFIRM)  
      --credential-context strings          Context name(s) in which to store credentials ($GPTSCRIPT_CREDENTIAL_CONTEXT)  
      --credential-override strings         Credentials to override (ex: --credential-override github.com/example/cred-tool:API_TOKEN=1234) ($GPTSCRIPT_CREDENTIAL_OVERRIDE)  
      --debug                               Enable debug logging ($GPTSCRIPT_DEBUG)  
      --debug-messages                      Enable logging of chat completion calls ($GPTSCRIPT_DEBUG_MESSAGES)  
      --default-model string                Default LLM model to use ($GPTSCRIPT_DEFAULT_MODEL) (default "gpt-4o")  
      --default-model-provider string       Default LLM model provider to use, this will override OpenAI settings ($GPTSCRIPT_DEFAULT_MODEL_PROVIDER)  
      --disable-cache                       Disable caching of LLM API responses ($GPTSCRIPT_DISABLE_CACHE)  
      --disable-tui                         Don't use chat TUI but instead verbose output ($GPTSCRIPT_DISABLE_TUI)  
      --dump-state string                   Dump the internal execution state to a file ($GPTSCRIPT_DUMP_STATE)  
      --events-stream-to string             Stream events to this location, could be a file descriptor/handle (e.g. fd://2), filename, or named pipe (e.g. \\.\pipe\my-pipe) ($GPTSCRIPT_EVENTS_STREAM_TO)  
      --force-chat                          Force an interactive chat session if even the top level tool is not a chat tool ($GPTSCRIPT_FORCE_CHAT)  
      --force-sequential                    Force parallel calls to run sequentially ($GPTSCRIPT_FORCE_SEQUENTIAL)  
      --github-enterprise-hostname string   The host name for a Github Enterprise instance to enable for remote loading ($GPTSCRIPT_GITHUB_ENTERPRISE_HOSTNAME)  
  -h, --help                                help for gptscript  
  -f, --input string                        Read input from a file ("-" for stdin) ($GPTSCRIPT_INPUT_FILE)  
      --list-models                         List the models available and exit ($GPTSCRIPT_LIST_MODELS)  
      --list-tools                          List built-in tools and exit ($GPTSCRIPT_LIST_TOOLS)  
      --no-trunc                            Do not truncate long log messages ($GPTSCRIPT_NO_TRUNC)  
      --openai-api-key string               OpenAI API KEY ($OPENAI_API_KEY)  
      --openai-base-url string              OpenAI base URL ($OPENAI_BASE_URL)  
      --openai-org-id string                OpenAI organization ID ($OPENAI_ORG_ID)  
  -o, --output string                       Save output to a file, or - for stdout ($GPTSCRIPT_OUTPUT)  
  -q, --quiet                               No output logging (set --quiet=false to force on even when there is no TTY) ($GPTSCRIPT_QUIET)  
      --save-chat-state-file string         A file to save the chat state to so that a conversation can be resumed with --chat-state ($GPTSCRIPT_SAVE_CHAT_STATE_FILE)  
      --sub-tool string                     Use tool of this name, not the first tool in file ($GPTSCRIPT_SUB_TOOL)  
      --system-tools-dir string             Directory that contains system managed tool for which GPTScript will not manage the runtime ($GPTSCRIPT_SYSTEM_TOOLS_DIR)  
      --ui                                  Launch the UI ($GPTSCRIPT_UI)  
      --workspace string                    Directory to use for the workspace, if specified it will not be deleted on exit ($GPTSCRIPT_WORKSPACE)
```

--------------------------------

### Execute GPTScript with Input Filters

Source: https://docs.gptscript.ai/tools/input-output-filters

Command to run the script and verify the filtered output.

```bash
gptscript script.gpt '{"message":"apple is great"}'  
  
# Then view the output:  
cat gptscript_output.txt
```

--------------------------------

### Inherited options

Source: https://docs.gptscript.ai/command-line-reference/gptscript_credential_show

Options inherited from parent commands for credential context.

```bash
--credential-context strings   Context name(s) in which to store credentials ($GPTSCRIPT_CREDENTIAL_CONTEXT)
```

--------------------------------

### Set a custom workspace directory

Source: https://docs.gptscript.ai/faqs

Override the default temporary workspace directory using the --workspace flag.

```bash
gptscript --workspace . my-script.gpt  
```

--------------------------------

### Credential delete options

Source: https://docs.gptscript.ai/command-line-reference/gptscript_credential_delete

Help flag available for the delete command.

```bash
  -h, --help   help for delete
```

--------------------------------

### Resolve Server Variables

Source: https://docs.gptscript.ai/tools/openapi

Demonstrates how GPTScript resolves variables within the server URL definition.

```yaml
servers:  
  - url: '{server}/v1'  
    variables:  
      server:  
        default: https://api.example.com  
```

--------------------------------

### Reference a Claude model in a script

Source: https://docs.gptscript.ai/alternative-model-providers

Use the model directive to specify a provider repository for Claude models.

```text
model: claude-3-haiku-20240307 from github.com/gptscript-ai/claude3-anthropic-provider  
  
Say hello world  
```

--------------------------------

### Credential override via environment variables

Source: https://docs.gptscript.ai/tools/credential-tools

Maps existing environment variables to a credential tool. Avoid this if multiple tools share the same variable names.

```text
toolA:ENV_VAR_1,ENV_VAR_2
```

--------------------------------

### Define Input Filters in GPTScript

Source: https://docs.gptscript.ai/tools/input-output-filters

Demonstrates chaining input filters to modify a message parameter through multiple tool calls.

```gptscript
# File name: script.gpt  
Param: message: the message from the user  
Tools: subtool  
Input Filter: appleToOrange  
  
Take the message and give it to the subtool. Then say "Done".  
  
---
Name: subtool  
Param: message: the message from the user  
Input Filter: orangeToBanana  
  
#!python3  
  
import os  
  
message = os.getenv("message", "")  
with open("gptscript_output.txt", "w") as f:  
    f.write(message)  
  
---
Name: appleToOrange  
  
#!python3  
  
import os  
  
def output(input: str):
    return input.replace("apple", "orange")  
  
print(output(os.getenv("INPUT", "")))  
  
---
Name: orangeToBanana  
  
#!python3  
  
import os  
  
def output(input: str):
    return input.replace("orange", "banana")  
      
print(output(os.getenv("INPUT", "")))
```

--------------------------------

### Reference an external Node.js script

Source: https://docs.gptscript.ai/tools/code-tool-guidelines

Use for complex Node.js tools requiring external dependencies.

```text
Name: my-node-tool  
  
#!/usr/bin/env node ${GPTSCRIPT_TOOL_DIR}/tool.js  
```

--------------------------------

### Save and resume chat state

Source: https://docs.gptscript.ai/faqs

Use these commands to persist chat history to a file and resume it in a later session.

```bash
gptscript --save-chat-state-file chat-state.json my-script.gpt  
```

```bash
gptscript --chat-state chat-state.json my-script.gpt  
```

--------------------------------

### Define an inline Python tool

Source: https://docs.gptscript.ai/tools/code-tool-guidelines

Use for simple Python scripts without external dependencies.

```text
Name: my-python-tool  
  
#!python3  
  
print('hello world')  
```

--------------------------------

### Reference an external Python script

Source: https://docs.gptscript.ai/tools/code-tool-guidelines

Use for complex Python tools requiring external dependencies.

```text
Name: my-python-tool  
  
#!/usr/bin/env python3 ${GPTSCRIPT_TOOL_DIR}/tool.py  
```

--------------------------------

### Use context in a tool

Source: https://docs.gptscript.ai/tools/context

Reference a context file within a GPTScript tool definition.

```gptscript
Context: ./my-context.txt  
Tools: sys.exec, sys.write  
  
Which processes on my system are using the most memory? Write their PIDs to a file called pids.txt.  

```

```gptscript
Context: ./my-context.txt  
Tools: sys.exec  
  
Which file in my current directory is the largest?  

```

--------------------------------

### gptscript parse options

Source: https://docs.gptscript.ai/command-line-reference/gptscript_parse

Specific flags available for the parse command.

```bash
  -h, --help           help for parse  
  -p, --pretty-print   Indent the json output ($GPTSCRIPT_PARSE_PRETTY_PRINT)
```

--------------------------------

### gptscript parse command usage

Source: https://docs.gptscript.ai/command-line-reference/gptscript_parse

Basic syntax for invoking the parse command.

```bash
gptscript parse [flags]
```

--------------------------------

### Accessing Credential Expiration in a Tool

Source: https://docs.gptscript.ai/tools/credential-tools

Demonstrates how to retrieve the nearest credential expiration time using the GPTSCRIPT_CREDENTIAL_EXPIRATION environment variable within a tool.

```python
Credential: my-credential-tool.gpt as myCred  
  
#!python3  
  
import os  
  
print("myCred expires at " + os.getenv("GPTSCRIPT_CREDENTIAL_EXPIRATION", ""))
```

--------------------------------

### Inspect OpenAPI operation IDs

Source: https://docs.gptscript.ai/examples/api

Identify available tool names by searching for operationId fields within the downloaded YAML file.

```bash
grep operationId openapi.yaml  
# ...  
# operationId: domains_delete_record  
# operationId: droplets_list  
# operationId: droplets_create  
# operationId: droplets_destroy_byTag  
# operationId: droplets_get  
# operationId: droplets_destroy  
# operationId: droplets_list_backups  
# operationId: droplets_list_snapshots  
# operationId: dropletActions_list  
# operationId: dropletActions_post  
# operationId: dropletActions_post_byTag  
# operationId: dropletActions_get  
# operationId: droplets_list_kernels  
# operationId: droplets_list_firewalls  
# operationId: droplets_list_neighbors  
# ...
```

--------------------------------

### Set workspace directory via CLI

Source: https://docs.gptscript.ai/tools/workspace

Configures the current directory as the workspace for a GPTScript execution.

```bash
gptscript --workspace . my-script.gpt
```

--------------------------------

### Set Credential Context via CLI

Source: https://docs.gptscript.ai/tools/credential-tools

Uses the --credential-context flag to specify a namespace for stored credentials during script execution.

```bash
gptscript --credential-context my-azure-workspace my-azure-script.gpt
```

--------------------------------

### Define a GPTScript workflow

Source: https://docs.gptscript.ai/examples/workflow

A complete GPTScript file defining the main task and supporting tools for tweet summarization.

```gptscript
Description: Summarize tweets  
Tools: sys.write, summarize-tweet  
  
Always summarize tweets synchronously in the order they appear.  
Never summarize tweets in parallel.  
Write all the summaries for the tweets at the following URLs to `tweets.md`:  
- https://x.com/acornlabs/status/1798063732394000559  
- https://x.com/acornlabs/status/1797998244447900084  
  
---  
Name: summarize-tweet  
Description: Returns a brief summary of a tweet in markdown  
Tools: get-hyperlinks, summarize-hyperlink, github.com/gptscript-ai/browser  
Parameters: url: Absolute URL of the tweet to summarize  
  
Return a markdown formatted summary of the tweet at ${url} using the following format:  
  
## <tweet-date> [<tweet-subject>](${url})  
  
Short summary of the tweet.  
  
### References  
  
<list of hyperlinks found in the tweet's text and their respective summaries>  
  
---  
Name: get-hyperlinks  
Description: Returns the list of hyperlinks in a given text  
Parameters: text: Text to extract hyperlinks from  
  
Return the list of hyperlinks found in ${text}. If ${text} contains no hyperlinks, return an empty list.  
  
---  
Name: summarize-hyperlink  
Description: Returns a summary of the page at a given hyperlink URL  
Tools: github.com/gptscript-ai/browser  
Parameters: url: HTTPS URL of a page to summarize  
  
Briefly summarize the page at ${url}.
```

--------------------------------

### Delete credential command syntax

Source: https://docs.gptscript.ai/command-line-reference/gptscript_credential_delete

Basic usage pattern for deleting a credential by name.

```bash
gptscript credential delete <credential name> [flags]
```

--------------------------------

### Define hyperlink processing tools

Source: https://docs.gptscript.ai/examples/workflow

Defines helper tools for extracting hyperlinks from text and summarizing the content found at those URLs.

```yaml
---
Name: get-hyperlinks
Description: Returns the list of hyperlinks in a given text
Parameters: text: Text to extract hyperlinks from

Return the list of hyperlinks found in ${text}. If ${text} contains no hyperlinks, return an empty list.

---
Name: summarize-hyperlink
Description: Returns a summary of the page at a given hyperlink URL
Tools: github.com/gptscript-ai/browser
Parameters: url: HTTPS URL of a page to summarize

Briefly summarize the page at ${url}.
```

--------------------------------

### Define Data Processing Assistant

Source: https://docs.gptscript.ai/examples/local-files

The configuration for the data processing assistant, including tools and workspace context.

```gptscript
Name: Data Processing Assitant
Description: An assistant to help you with processing data found in files on your workstation. Helpful for querying spreadsheets, CSVs, JSON files, and PDFs.
Tools: github.com/gptscript-ai/structured-data-querier, github.com/gptscript-ai/pdf-reader
Context: github.com/gptscript-ai/context/workspace
Chat: true

You are a helpful data processing assistant. Your goal is to help the user with data processing.
Help the user accomplish their tasks using the tools you have.
When the user starts this chat, say hi, introduce yourself, and ask what you can help with.
```

--------------------------------

### gptscript getenv

Source: https://docs.gptscript.ai/command-line-reference/gptscript_getenv

Retrieves the value of an environment variable, optionally providing a default value if the variable is not set.

```APIDOC
## gptscript getenv

### Description
Looks up an environment variable for use in GPTScript tools. If the variable is not found, the provided default value is returned.

### Usage
`gptscript getenv [flags] KEY [DEFAULT]`

### Parameters
- **KEY** (string) - Required - The name of the environment variable to retrieve.
- **DEFAULT** (string) - Optional - The value to return if the environment variable is not set.

### Options
- **-h, --help** - Displays help information for the getenv command.
```

--------------------------------

### gptscript parse

Source: https://docs.gptscript.ai/command-line-reference/gptscript_parse

Parses a GPTScript file and outputs the result. This command supports flags for formatting the output.

```APIDOC
## gptscript parse

### Description
Parses a GPTScript file and returns the parsed structure in JSON format.

### Flags
- **-p, --pretty-print** (boolean) - Optional - Indent the JSON output.
- **-h, --help** (boolean) - Optional - Display help for the parse command.
```