Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.perplexity.ai/llms.txt

Use this file to discover all available pages before exploring further.

This page tracks upcoming and in-progress work only. Once a feature ships, it is removed from this roadmap and documented in the changelog.

​
Upcoming and In Progress

We are introducing a Sandbox API for isolated environments for executing Python and Bash code. Each sandbox runs in its own container with dedicated resources, supporting file operations, background processes, and state persistence via pause/resume.
  • Dedicated per-sandbox containers for stronger isolation
  • Python and Bash execution with file and process support
  • Pause/resume state persistence for iterative workflows
We are rolling out performance improvements while preserving compatibility.
  • Same endpoint contracts with backward compatibility
  • Latency improvements for common request patterns
  • Automatic rollout with no migration required for most users
We’re expanding multimodal support to include direct video uploads.
  • Video content analysis for uploaded files
  • Frame-level reasoning for time-specific insights
  • Visual scene understanding across longer media
  • Multimodal retrieval across text, image, and video context
We’re working on deeper file and data-source access for API organizations.
  • Repository and connector search across organization data
  • Multi-format support across common document types
  • External data source integration for enterprise workflows
We plan to add webhook callbacks for async workflows so long-running jobs can notify your systems automatically.
  • Job completion webhooks for async requests
  • Failure webhooks with retry-safe delivery semantics
  • Signature verification for secure callback handling
We will continue to expand model coverage and improve SDK ergonomics.
  • Additional model releases and lifecycle clarity
  • SDK quality-of-life improvements for Python and TypeScript
  • More production-ready examples for common integration patterns
We’re addressing context persistence limits by improving memory-oriented workflows.
  • Session-aware state handling
  • Reduced manual context stitching across requests
  • Improved control over long-running conversational workflows
We are improving how failures are reported and diagnosed across APIs.
  • More actionable error messages
  • Stronger remediation guidance in docs
  • Clearer retry and backoff recommendations
We’re building voice-native interaction capabilities for real-time experiences.
  • Direct voice input and streamed audio output
  • Multi-language support
  • Low-latency conversation loops
We’re building a dedicated API console — a standalone interface for developers that is separate from the Perplexity website. It will serve as the primary hub for API key management, usage visibility, and team controls.
  • Standalone from the Perplexity app — purpose-built for API developers
  • API key management with fine-grained permissions and rotation
  • Usage and cost visibility at the key and team level
  • Organization controls for team access and collaboration
We’re expanding analytics visibility for API teams.
  • Query and latency analytics
  • Error trend monitoring
  • Cost and usage forecasting
We’re expanding financial research capabilities with richer structured access.
  • Market and filing workflows
  • Better controls for finance-specific retrieval
  • More end-to-end examples for analyst use cases
Documentation improvements remain in progress.
  • Clearer API selection guides
  • More opinionated implementation guides
  • Higher-fidelity, production-oriented examples
We intend to bring selected experimental features into developer previews.
  • Early access feature channels
  • Prototype integrations for feedback
  • Faster developer feedback loops for roadmap planning
For shipped updates and release dates, see the changelog.