Paxalabs MCP server
{
"mcpServers": {
"paxalabs": {
"command": "npx",
"args": [
"-y",
"@paxalabs/mcp@latest"
],
"env": {
"PAXA_API_KEY": "<paxa_api_key>",
"PAXA_BASE_URL": "https://..."
}
}
}
}{
"mcpServers": {
"paxalabs": {
"command": "npx",
"args": [
"-y",
"@paxalabs/mcp@latest"
],
"env": {
"PAXA_API_KEY": "<paxa_api_key>",
"PAXA_BASE_URL": "https://..."
}
}
}
}{
"mcpServers": {
"paxalabs": {
"command": "npx",
"args": [
"-y",
"@paxalabs/mcp@latest"
],
"env": {
"PAXA_API_KEY": "<paxa_api_key>",
"PAXA_BASE_URL": "https://..."
}
}
}
}{
"servers": {
"paxalabs": {
"command": "npx",
"args": [
"-y",
"@paxalabs/mcp@latest"
],
"env": {
"PAXA_API_KEY": "${input:paxa_api_key}",
"PAXA_BASE_URL": "${input:paxa_base_url}"
}
}
},
"inputs": [
{
"id": "paxa_api_key",
"type": "promptString",
"description": "Paxa api key",
"password": true
},
{
"id": "paxa_base_url",
"type": "promptString",
"description": "Paxa base url",
"password": true
}
]
}{
"mcpServers": {
"paxalabs": {
"command": "npx",
"args": [
"-y",
"@paxalabs/mcp@latest"
],
"env": {
"PAXA_API_KEY": "<paxa_api_key>",
"PAXA_BASE_URL": "https://..."
}
}
}
}file .mcp.json (in your project root)file claude_desktop_config.jsonfile ~/.cursor/mcp.jsonfile .vscode/mcp.jsonfile ~/.codeium/windsurf/mcp_config.jsoninstall npx -y @paxalabs/mcpversion 0.1.0-beta.10
What it does
MCP server for Thai and English speech AI from Paxa Labs. Provides text-to-speech with local playback, translation to Thai, and OCR for PDFs and images. Designed for agents that need to speak, listen, and process documents.
- Have an agent read long-form content aloud through managed playback queues while synthesizing ahead to avoid gaps.
- Translate text into Thai with controls for formality, glossary, and context.
- Extract text from PDFs, images, and screenshots using OCR and convert to Markdown or structured blocks.
Tools this server exposes
| Tool | Does |
|---|---|
| speak | Synthesize and play a short line through speakers, blocking until playback finishes. |
| queue_speech | Queue long content for playback with automatic chunking and ahead-of-time synthesis to avoid gaps. |
| control_playback | Control the audio queue: check status, pause, resume, skip, or clear. |
| play_audio | Play a local audio file through the speakers. |
| text_to_speech | Synthesize speech to an audio file (mp3, opus, wav) without playing it. |
| translate_to_thai | Translate any language into Thai with controls for formality, glossary, and context. |
| ocr_document | Extract text from a local PDF, PNG, JPEG, or WebP image into Markdown or structured blocks. |
| list_voices | List available TTS voices with character descriptions. |
| list_models | List available models with limits and pricing information. |
| get_account | Retrieve credit balance, plan details, and rate limits. |
Before you enable it
An MCP server runs with whatever access you give it. Start with a development account rather than production, and check which tools the server exposes before you let an agent call them unattended.