Give an AI agent Perplexity as a tool and it can look things up for itself instead of guessing. It can search the web, run cited chat completions, pull academic papers, and search SEC filings mid-task, then feed the answers straight back into whatever it is working on. When it needs a clean object rather than prose it can request structured JSON output, and it can generate embeddings to compare or rank text.
Pick the operations you want and each becomes a tool the agent can call directly while it works.
Generate a search-grounded model answer for a chat conversation.
Submit a long-running (e.g. deep-research) completion as a background job.
List all async chat-completion jobs for the account.
Poll the status and retrieve a specific async completion job.
Get raw, ranked web search results (title, URL, snippet).
Run a frontier model with Perplexity web-search tools and filters.
Generate vector embeddings for text.
Generate context-aware embeddings (document + surrounding context).
List the models available to the account.
Programmatically create a new API key/token for the account.
Revoke an existing API key/token (rotation).
Retrieve account usage/analytics data.
Chat completion grounded in scholarly sources (search_mode: academic).
Chat completion grounded in SEC filings (search_mode: sec).
Chat completion with a JSON schema response_format for validated output.
OpenCode is an open, model-agnostic command-line coding agent. In NoClick it runs hosted and gains your apps as tools, so it can act on real systems rather than just edit files. Wire an integration into the agent and OpenCode can use that app’s operations directly while it works. It is a flexible choice when you want an open agent with broad model support and real tool access.
NoClick runs OpenCode for you with Perplexity wired in as tools. Connect your account and run.