OpenAI launches Agents API public beta for long-running cloud agents
OpenAI announced the Agents API in public beta, a developer platform that lets users build cloud‑based AI agents capable of running for days, handling tool calls, and managing context across multiple ranges. The API adds features such as Tool Search to load only needed tool definitions, Programmatic Tool Calling for parallel operations, and the ability to split work among sub‑agents in separate…
What you can do with it
AI at Work →Draft articles from official sources only with ChatGPT
What you get You can create estimates using approved price lists and reply to inquiries.
Separates generation.
- Who for
- marketers, founders, support teams and operations
- Cost
- free tier
- Effort
- an afternoon
Use it for
- Draft articles from official sources only
- Create estimates using approved price lists
- Reply to inquiries
How to set it up · From the article
- Pick one reversible task (drafting, summarizing, formatting)
- Write down the current human steps and expert judgments
- Restrict AI to a few approved sources
- Use separate prompts for generation and verification
Watch out Requires manual setup of sources, templates, and checklists
ChatGPT in the tool directory · 5 changes covered
Key points from the news
- OpenAI released the Agents API in public beta, enabling developers to build long‑running cloud agents with tool search and programmatic tool calling.
- OpenAI introduced ChatGPT for Financial Services, built on GPT‑6 Astra, integrating data from over 50 connectors and offering enterprise security features.
- Anthropic published a threat‑intelligence report covering AI misuse cases from Dec 2025 to Aug 2026, highlighting AI‑driven cyber and influence operations.
On the same day OpenAI introduced ChatGPT for Financial Services, powered by GPT‑6 Astra. The product integrates data from sources like Daloopa, PitchBook, LSEG News and Crunchbase, offers over 50 connectors, and provides enterprise security controls such as SAML SSO, role‑based access and encrypted storage. It targets investment banking and equity research, with template‑driven document generation for Excel, Word and PowerPoint.
Anthropic released a threat‑intelligence report covering AI misuse from December 2025 to August 2026. The report highlights AI‑driven cyber operations, influence campaigns, and other threats, noting that AI now coordinates reconnaissance and intrusion tasks. Anthropic cautions that attackers can modify tools quickly with AI assistance, urging stronger permission checks, logging and human approval.
Model page: GPT-6 Astra →
The story so far
4 episodes →- OpenAI launches Agents API public beta for long-running cloud agentsthis story
[September 10] Latest AI News | Agents API, ChatGPT for Finance, and AI Abuse Reports Explained
note.com · 21 September 2026
Loading the full article…
This text was published by note.com and written by AI情報屋 sho｜AI実践・検証. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
Coverage and discussion
2sourcesThe headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Agents & Tools
All →- Google launches Gemini 3.8 Live voice model for developers · 5 src
- Meta's Muse AI agent tops US App Store, drives stock surge · 15 src
- Trane reduces HVAC diagnostic time from 20 minutes to 20 seconds with Bedrock AgentCore · 1 src
- AstroForge plans to fly AI‑controlled spacecraft Solo in 2027 · 1 src
- Brave launches privacy‑first AI assistant Leo · 1 src
Comments
via GitHub Discussions