API Updates

API changes, new endpoints, and developer tools.

June 2026

Anthropicapi

Link·Jun 17, 2026

Claude Workload Identity Federation

A way for companies to connect to the Claude API without storing permanent API keys. Each request gets a short-lived credential issued on the spot. Enterprise and developer setups only.

Best for: Security teams that do not want long-lived Claude API keys sitting in their systems.
When to use it: You are wiring Claude into AWS, Google Cloud, or GitHub Actions and want those systems to authenticate without a stored key.
Who benefits: Engineering and security teams running Claude in production, not end users.

OpenAIapi

Link·Jun 9, 2026

Image Results in Responses API Web Search

Apps built on OpenAI's Responses API can now pull image results from web search, not just text and links.

Best for: Apps that need to show products, places, or visual references, not just describe them.
When to use it: Your app's users expect to see images in search results rather than read about them.
Who benefits: Developers building search, shopping, or inspiration features on the Responses API.

OpenAIapi

Link·Jun 4, 2026

Moderation Scores in the Responses and Completions API

OpenAI's API can now return content-moderation signals in the same request that generates a response.

Best for: teams that need to screen AI inputs or outputs without making a second API call
When to use it: you want moderation signals returned next to the generated text so you can act in one step
Who benefits: developers building apps that handle user-generated or sensitive content

Perplexityapi

Link·Jun 1, 2026

Search as Code

A search setup for developers building agents on Perplexity: the agent writes its own Python to run many searches at once instead of one at a time. Now the default in Perplexity Computer.

Best for: Agents that need to pull from and sift through many sources in a single task, like deep research or data gathering.
When to use it: Building on the Perplexity Agent API and you want the agent to run many searches at once instead of looping one query at a time.
Who benefits: Developers and engineering teams building AI research or agent tools on Perplexity.

May 2026

Googleapi

Link·May 19, 2026

Managed Agents in the Gemini API

New managed agents capability for developers building on Gemini.

Best for: Building production agents on Gemini without running your own agent runtime
When to use it: You want Google to handle agent orchestration, state, and execution
Who benefits: Developers and platform teams building on the Gemini API

Anthropicapi

Link·May 11, 2026

Claude Platform on AWS (Generally Available)

Anthropic's full Claude API is now available directly on AWS, with AWS authentication and billing.

Best for: AWS-native teams that want the latest Claude API features without going through Bedrock.
When to use it: you want to use new Claude features and prefer AWS billing.
Who benefits: enterprise developers building on AWS who've been stuck waiting for Bedrock feature parity.

Perplexityapi

Link·May 6, 2026

Finance Search in Perplexity Agent API

One API call returns licensed financial datasets, real-time market data, and cited web sources for finance agents.

Best for: Building finance agents that need verifiable, current market data without stitching together multiple data providers.
When to use it: Agents need live market data (prices, fundamentals, earnings, filings) with citations from a single tool call.
Who benefits: Fintech, research, and analyst tooling teams who previously integrated separate market data feeds and search providers.

April 2026

Googleapi

Link·Apr 29, 2026

Google-Managed MCP Servers (50+ Services)

Google Cloud now offers 50+ fully managed Model Context Protocol servers for connecting AI agents to its services.

Best for: building agents that need to read or act on Google Cloud and Workspace data
When to use it: wiring Gemini CLI, Claude, ChatGPT, VS Code, LangChain, or ADK agents into BigQuery, Spanner, GKE, Drive, Calendar, and similar services
Who benefits: developers and platform engineers running multi-tool agent workflows on Google Cloud

OpenAIapi

Link·Apr 28, 2026

OpenAI on AWS

OpenAI models, Codex, and Managed Agents are now available inside Amazon Bedrock in limited preview.

Best for: enterprises already on AWS that want OpenAI models without standing up a separate vendor relationship.
When to use it: your team's spend, security review, and orchestration already live in Bedrock.
Who benefits: AWS-native engineering and platform teams whose procurement has blocked direct OpenAI deals.

OpenAIapi

Link·Apr 24, 2026

GPT-5.5 in the API

GPT-5.5 and GPT-5.5 Pro are now available to developers through the OpenAI API.

Best for: developers building tools and agents that need OpenAI's newest agentic model
When to use it: you want to wire GPT-5.5's long-task and tool-use behavior into your own product or workflow
Who benefits: developers on the OpenAI API