246a5429af
- Add 'Passing API Keys to Tools' section to registry-execute README - Add 'Understanding requiredEnvVars' section to registry-search README - Include example of pre-configuring keys for agents - Document tools that don't require keys - Bump both packages to 0.1.2
125 lines
3.2 KiB
Markdown
125 lines
3.2 KiB
Markdown
# @tpmjs/registry-search
|
|
|
|
Search the TPMJS tool registry from any AI SDK agent. Discover thousands of tools dynamically.
|
|
|
|
## Installation
|
|
|
|
```bash
|
|
npm install @tpmjs/registry-search
|
|
# or
|
|
pnpm add @tpmjs/registry-search
|
|
```
|
|
|
|
## Usage
|
|
|
|
```typescript
|
|
import { streamText } from 'ai';
|
|
import { anthropic } from '@ai-sdk/anthropic';
|
|
import { registrySearchTool } from '@tpmjs/registry-search';
|
|
import { registryExecuteTool } from '@tpmjs/registry-execute';
|
|
|
|
const result = streamText({
|
|
model: anthropic('claude-sonnet-4-20250514'),
|
|
tools: {
|
|
registrySearch: registrySearchTool,
|
|
registryExecute: registryExecuteTool,
|
|
},
|
|
system: `You have access to the TPMJS tool registry.
|
|
Use registrySearch to find tools, then registryExecute to run them.`,
|
|
prompt: 'Search for web scraping tools and scrape https://example.com',
|
|
});
|
|
|
|
// The agent can now:
|
|
// 1. Search for tools: registrySearch({ query: "web scraping" })
|
|
// 2. Execute found tools: registryExecute({ toolId: "@firecrawl/ai-sdk::scrapeTool", params: { url: "..." } })
|
|
```
|
|
|
|
## Tool: registrySearchTool
|
|
|
|
Search the TPMJS registry to find AI SDK tools.
|
|
|
|
### Parameters
|
|
|
|
| Name | Type | Required | Description |
|
|
|------|------|----------|-------------|
|
|
| `query` | string | Yes | Search query (keywords, tool names, descriptions) |
|
|
| `category` | string | No | Filter by category |
|
|
| `limit` | number | No | Max results (1-20, default 5) |
|
|
|
|
### Categories
|
|
|
|
- `web-scraping`
|
|
- `data-processing`
|
|
- `file-operations`
|
|
- `communication`
|
|
- `database`
|
|
- `api-integration`
|
|
- `image-processing`
|
|
- `text-analysis`
|
|
- `automation`
|
|
- `ai-ml`
|
|
- `security`
|
|
- `monitoring`
|
|
|
|
### Returns
|
|
|
|
```json
|
|
{
|
|
"query": "web scraping",
|
|
"matchCount": 3,
|
|
"tools": [
|
|
{
|
|
"toolId": "@firecrawl/ai-sdk::scrapeTool",
|
|
"name": "scrapeTool",
|
|
"package": "@firecrawl/ai-sdk",
|
|
"description": "Scrape any website into clean markdown",
|
|
"category": "web-scraping",
|
|
"requiredEnvVars": ["FIRECRAWL_API_KEY"],
|
|
"healthStatus": "HEALTHY",
|
|
"qualityScore": 0.9
|
|
}
|
|
]
|
|
}
|
|
```
|
|
|
|
### Understanding `requiredEnvVars`
|
|
|
|
The `requiredEnvVars` field tells you which API keys a tool needs to function. When executing a tool with `@tpmjs/registry-execute`, pass these keys in the `env` parameter:
|
|
|
|
```typescript
|
|
// 1. Search finds a tool that needs FIRECRAWL_API_KEY
|
|
const searchResult = await registrySearchTool.execute({ query: 'web scraping' });
|
|
// searchResult.tools[0].requiredEnvVars = ["FIRECRAWL_API_KEY"]
|
|
|
|
// 2. Execute the tool with the required key
|
|
const result = await registryExecuteTool.execute({
|
|
toolId: '@firecrawl/ai-sdk::scrapeTool',
|
|
params: { url: 'https://example.com' },
|
|
env: { FIRECRAWL_API_KEY: 'fc-xxx' } // Pass required keys here
|
|
});
|
|
```
|
|
|
|
Tools with an empty `requiredEnvVars` array don't need any API keys.
|
|
|
|
## Environment Variables
|
|
|
|
| Variable | Default | Description |
|
|
|----------|---------|-------------|
|
|
| `TPMJS_API_URL` | `https://tpmjs.com` | Base URL for the registry API |
|
|
|
|
### Self-Hosted Registry
|
|
|
|
To use your own TPMJS registry:
|
|
|
|
```bash
|
|
export TPMJS_API_URL=https://registry.mycompany.com
|
|
```
|
|
|
|
## Related
|
|
|
|
- [@tpmjs/registry-execute](https://www.npmjs.com/package/@tpmjs/registry-execute) - Execute tools found with this package
|
|
- [TPMJS Registry](https://tpmjs.com) - Browse all available tools
|
|
|
|
## License
|
|
|
|
MIT
|