GPT Actions Compendium - List of Free and Open Source GPT Actions

Here is a GitHub repository with several ready-to-use schemas for actions you can add to your GPTs.

23 Likes

This is a great topic!
Would be even better if we can expand it with a collection of community actions to share with all builders. If you reading this have a action to share we will find a way to add this, Iā€™m sure.

By the wayā€¦
How do I create an action to get a response from the chat.completions endpoint?
I am getting a ā€˜Invalid URLā€™ error, even though the address is straight from the API specification.

ā€œdomainā€: ā€œapi.openai.comā€,
ā€œmethodā€: ā€œpostā€,
ā€œpathā€: ā€œ/v1/chat/completionsā€,

The authentication method is set to API and Bearer but I am somewhat at a loss what I am missing hereā€¦

5 Likes

Here are some more resources

7 Likes

Can you post your schema or a link to it?

2 Likes

Yes, thanks for looking into this.

openapi: 3.0.0
info:
  title: OpenAI Chat Completions Forwarding API
  description: This API forwards messages to the OpenAI Chat Completions endpoint.
  version: 1.0.0
servers:
  - url: https://api.openai.com
    description: OpenAI Chat Completions API server
paths:
  /v1/chat/completions:
    post:
      operationId: forwardChatCompletion
      summary: Forward message to OpenAI Chat Completions
      description: Sends a request to the OpenAI Chat Completions endpoint with a specified model and message array.
      requestBody:
        required: true
        content:
          application/json:
            schema:
              type: object
              properties:
                model:
                  type: string
                  example: "gpt-3.5-turbo" or "gpt-4"
                  description: The model to be used for completion.
                messages:
                  type: array
                  description: An array of messages for the chat completion.
                  items:
                    type: object
                    properties:
                      role:
                        type: string
                        enum: [system, user]
                        description: The role of the message sender.
                      content:
                        type: string
                        description: The content of the message.
      responses:
        '200':
          description: Successful response with chat completion data
          content:
            application/json:
              schema:
                type: object
                properties:
                  id:
                    type: string
                    description: Unique identifier for the completion request.
                  object:
                    type: string
                    example: "text_completion"
                  created:
                    type: integer
                    format: int64
                    description: Timestamp of the creation time.
                  model:
                    type: string
                  choices:
                    type: array
                    items:
                      type: object
                      properties:
                        message:
                          type: string
                          description: The completion message from the model.

1 Like

Thanks! but Iā€™ve got error ā€˜There was an error generating a responseā€™ with ā€˜Regenerateā€™ button after ā€˜Allowā€™.
Arxiv and Klarna were tested.

Hi everyone, Iā€™m the creator of this repository and I just wanted to let everyone know that if you want to add your schema to it, all you have to do is open a new issue and fill out the form:

The rest (validation, approval, and adding it to the repoā€™s front page) is all taken care of automatically.

4 Likes

Hi @august.eight, would you mind telling me if thereā€™s a specific error message? I might be able to help.

1 Like

Hi @bapo2,

I was able to make both mentioned actions work by simply copy and pasting yesterday, but then forgot to post the results here in the forum.

Thanks for joining!

3 Likes

Thanks for the links! It is great to have the specs on hand :eyes:

Found a very detailed and seemingly complete OpenAI OpenAPI spec on GitHub.

5 Likes

Below is error message for Test regarding ā€˜https://raw.githubusercontent.com/bapo2/gpt-actions/main/schemas/apisguru-search/schema.jsonā€™.

[debug] Calling HTTP endpoint
{
  "domain": "api.apis.guru",
  "method": "get",
  "path": "/providers.json",
  "operation": "listProviders",
  "operation_hash": "7b3e029e1623ea780b0d81cd6e4b36a7bfa2c5b0",
  "is_consequential": false,
  "params": {}
}

This isnā€™t an error message.

Below is snapshot of mine for preview tab.

That could be anything from some sort of transient issue to the API endpoint blocking traffic from OpenAIā€™s servers to an authentication issue.

This sort of thing can be tricky to debug, especially if you donā€™t have direct access to either of the servers involved in the call.

I donā€™t have right now to experiment with it, but Iā€™ll try to look at it later.

Meantime, I also have being tried to solve my situation. And, it has been solved luckly.
My fault for the error was ā€˜saveā€™. I should save before conduct test.
test in preview page without save is not working.
Therefore, Save before test the Action is necessary.

Thanks for your efforts! @anon22939549 @bapo2

It was caused by my insufficient understanding for the system.

3 Likes

I went ahead and submitted Web Requestsā€™ (the one powering WebGPTšŸ¤–).

Our endpoints are currently free and OpenAIā€™s GPTā€™s IPā€™s are all whitelisted for free use. No authentication required.

1 Like

sometimes you just get flagged randomly for inexplicable reasons :confused:

Summary

The title of the topic is ā€œList of actionsā€.
These actions can be used by other builders to create their own GPTs.
This topic is not intended to promote oneā€™s own GPTs, unless they specifically help with building actions.

I think thatā€™s really simple.
@JD_2020 Then help me understand why the link you posted is not a promotion but a on-topic post.

Edit: I see it now! You created an issue(?) in the repo. I suppose thatā€™s all fine then.