Stay informed about Portkey's newest releases, features, and improvements
✨How are companies actually deploying LLMs in production?
We went beyond the hype and analyzed 2T+ tokens to understand the real patterns when deploying LLMs in production - from provider dynamics to performance metrics.
Join us today as we break down insights from 2T+ LLM tokens across 90+ regions.
We analyzed real-world production data through Portkey's AI Gateway to understand how companies are actually implementing LLMs at scale.
Today we're sharing key insights around:
📊 Provider adoption trends in 2025
✨ Real-world performance metrics & benchmarks
⚡ Production patterns & cost optimization
Join us for the LLMs in Prod'25 launch webinar today at 10:00 AM PT :
Dive into your AI journey: requests processed, response times, reliability stats, and the innovative ways your team used LLMs.
Plus, see how much you saved through cache optimization!
Check your Portkey associated email for the full story ✨
Portkey now has comprehensive support for Perplexity AI's most powerful features, making it easier to build more sophisticated AI applications with granular control over search and responses.
Search Domain Filter
Control knowledge sources for the LLM by whitelisting or blacklisting up to 3 domains for citations. Perfect for ensuring responses draw from trusted sources only.
Image Results Integration
Enable LLM responses to include relevant images, enriching the user experience with visual context alongside text responses.
Related Questions
Automatically generate follow-up questions to enhance user engagement and exploration of topics.
Search Recency
Filter search results by recency (hour/day/week/month) to ensure your AI always works with the most current information.
All features are available through simple parameter flags in both our Python and NodeJS SDKs.
Check Portkey docs for detailed implementation examples.
Portkey now supports flexible enterprise authentication options with the addition of SAML support alongside our existing OIDC integration. Orgs can now choose their preferred SSO protocol based on their infrastructure needs.
What's New:
If you are using identity providers like Okta, Azure AD, or Google Workspace - you can now configure your preferred authentication protocol through our SSO settings page.
This feature is a part of our Enterprise offering. Read more about it here.
Code blocks in LLM responses are now more readable and useful with dedicated language indicators. When the LLM includes code snippets using markdown syntax, the programming language is prominently displayed above each code block.
We also mapped out the whole MCP ecosystem put and all available MCP servers, from reference implementations to community contributions, in one place.
Browse here: https://portkey.ai/mcp-servers
Today, we're announcing MCP (Model Context Protocol) Client by Portkey — the world's first truly magical agent platform that turns "write me an app" into "here's your deployed app."
No complex setup. No integration hell. Just pure possibility.
Sign up for the waitlist here → https://portkey.ai/mcp
With Grok's next-gen models, you can now integrate cutting-edge AI capabilities, including:
Portkey's powerful features like reliability, caching, and security ensure smooth integration of Grok into your production-grade AI apps.
Check out the docs and get started now! ➡️ xAI
We've just shipped a console UI to the open source Gateway, to help you see all of your local logs in one place.
Just hit npx @portkey-ai/gateway and access the console on http://localhost:8787/public/