Skip to content
Log inSign up

The API toolkit for your agent.

Discover capabilities. Call APIs. Adapt to the results.
One CLI, in your terminal and your agent's toolbox.

npm install -g @apibasket/cli
Add to your agent Read CLI docs

API Basket CLI

Agent workflow

Agent activity

  1. Discover APIs — pending

    Describe your task to find API combinations.

  2. Select an API combination — pending

    Choose the combination that best fits your task.

  3. Compose API calls — pending

    Follow the skill document to compose CLI commands.

  4. Finish the task — pending

    Retrieve the result using the task ID.

@apibasket/cli

Demo ready.

apibasket search "turn my pdf to excel"
{
  "goal": "turn my pdf to excel",
  "proposals": [
    {
      "proposal": "Use OCR API and Excel Builder API",
      "proposal_steps": [
        {
          "exe": "apibasket call pdf_api_0x5067",
          "inputs": [
            "--params",
            "{\"input\":\"/path/to/your/file\"}",
            "--link-api",
            "excel_builder_api_0xa7a1",
            "--output",
            "/path/to/your/output_dir/output.xlsx"
          ]
        }
      ],
      "estimate_credit": 2
    }
  ]
}
apibasket use "Use OCR API and Excel Builder API"
# PDF to Excel

## Selected APIs
- OCR API: pdf_api_0x5067
- Excel Builder API: excel_builder_api_0xa7a1

## Required input
Set params.input to the PDF file path.

## Instructions
1. Call pdf_api_0x5067 with the input PDF.
2. Link excel_builder_api_0xa7a1 using --link-api.
3. Set --output to /path/to/your/output_dir/output.xlsx.
4. Read request_id and check its status before reporting completion.

Estimated credits: 2

Follow this skill to compose and execute the API call.
apibasket call pdf_api_0x5067 \
  --params '{"input":"/path/to/your/file"}' \
  --link-api excel_builder_api_0xa7a1 \
  --output /path/to/your/output_dir/output.xlsx
{
  "request_id": "req_0x59f5_YW1hemlu_YXBpYmFza2V0",
  "credits_reserved": 2
}
apibasket status req_0x59f5_YW1hemlu_YXBpYmFza2V0 --json
{
  "request_id": "req_0x59f5_YW1hemlu_YXBpYmFza2V0",
  "status": "completed",
  "data": {
    "download_url": "https://media.apibasket.com/your_account/YW1hemlu_YXBpYmFza2V0"
  }
}

Proposed CLI syntax and simulated results.

Give your agent the instructions.
Let it choose the steps.

The LLM decides what to do. API Basket CLI provides the tools.

  1. Install the CLI

    Available to your agent's shell.

    npm install -g @apibasket/cli

  2. Connect API Basket

    Authenticate once in your environment.

    apibasket login

  3. Add the agent skill

    Give your agent API discovery and usage instructions.

SKILL.md

---
name: apibasket
description: Discover and call APIs for a user goal.
---
 
# Working with API Basket
1. Discover APIs: apibasket search "<user goal>"
2. Compare the proposed API combinations and estimated credits.
3. Select a proposal: apibasket use "<proposal>"
4. Follow the returned skill to compose API calls.
5. Read the output and decide the next action.
6. On retryable errors, follow retry_after; retry at most twice.
7. Check the request status, then answer the user.

Works through the agent's shell execution tool.

Built for the next decision.

search · proposals

Discover at runtime

The agent finds APIs and compares combinations for the task.

use · skill

Compose with context

Returned instructions guide the agent's next API call.

request_id · status

Verify the outcome

Check completion before reporting the result.

Design proposal · CLI commands, skill format and results are illustrative.