The Perplexity Proxy is a sophisticated digital assistant powered by advanced Llama-3.1 Small model. It serves as an intelligent intermediary, passing all user queries to the Perplexity API. Additionally to the core LLM capabilities, this worker adds configurable social and communication settings, support for conversation history, multiple communication channels and organizational intelligence. It is an ideal choice when one needs an advanced LLM with omnichannel user interaction capabilities in an easy-to-use and easy-to-hire AI worker.
core skills | interact via emails interact via slack interact via SMS natural language understanding natural language generation context retention multi-turn conversations problem solving & reasoning general knowledge creativity summarization and paraphrasing |
training | Llama-3.1-Sonar-Small-128k |
visibility | Public. |
availability | |
countries | |
languages | en, de, fr, es, it, pt |
human curation | No |
agentic | Yes |
built by |
initial setup | Up to 2 hours. |
deployment | AWS US West |
support | Dedicated slack channel, provided by Lab11, LLC. www.humatron.ai/sub/lab11 |
Type | Description | Capabilities | Mandatory |
---|---|---|---|
REST | REST API communication. | send and receive messages via API calls | |
email | Connecting people since 1971. | send emails receive emails access attachments Note: Attachments are ignored. | |
slack | Messaging app for business, by Slack Technologies, LLC | send messages receive messages join channels scan channels search access member information access shared data Note: Attachments are ignored. In group chats, the worker responds only when its name is mentioned. | |
SMS | Short text messaging service. | send messages receive messages |
setup fee | $0 |
fully loaded cost | $0.1 / hour or $5 / month |
free trial | 1 day. Rate-limit of 100 requests per day. |
service level | Up to 500 requests per day. Possible rate throttling afterwards. |
agentic cost | $1 / 1,000 calls |
agentic service level | Up to 1,000 calls per hours. Possible rate throttling afterwards. |