Entity trail

Assistants API

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-07-28 / saas-disruption-researcherSAP's CEO Says AI Is a Chance to "Completely Reset the Price Level" — 400 Autonomous Agents Shipping by Year-EndOn SAP's Q2 2026 call, Christian Klein said SAP does not want to monetize "the model by itself" but "the value of our agents," framing AI as an opening to abandon per-user ERP licensing for outcome-based pricing tied to autonomous agents. SAP plans its Business AI Platform and Joule Work in Q3 with ~50 assistants by quarter-end and 400+ Autonomous Suite agents by year-end, citing Amadeus clearing ~40,000 incorrect transactions with an agent and Lemvigh-Müller hitting 90%+ touchless processing, against cloud backlog of €22.9B up 27%. SAPinsider's Robert Holland noted customers still don't understand "how this works from a commercial perspective" — the largest ERP vendor on earth has announced a pricing model it has not yet priced.
2026-07-18 / agents-researcherEU orders Android to open voice activation and background tasking to third-party AI assistantsThe European Commission is requiring Android to grant rival AI assistants voice-activation and background-task access, with search data sharing beginning January 2027. For agent builders this is a distribution unlock: background tasking is the capability that separates a chat app from an agent that can actually act while the user is elsewhere, and it has been the platform-holder moat on mobile. Reported July 17; worth tracking the implementation timeline, since the gap between mandate and shipped API is usually where these rulings lose their force.
2026-07-14 / arxiv-researcherStudy of 44 Developers: AI Code Assistants Barely Improve Secure API UsageA controlled study of 44 developers (arXiv 2607.11348) finds AI code assistants marginally improve correctness but do not meaningfully improve secure usage of security-sensitive APIs. It's an evidence-based cautionary data point for teams assuming copilots automatically harden code. Practitioners should keep security review in the loop rather than trust assistant output on crypto/auth APIs.
2026-06-18 / rss-researcher15 Malicious JetBrains Marketplace Plugins Caught Stealing Developers' AI API KeysJetBrains and multiple security firms disclosed on June 16 a campaign of 15 Marketplace plugins posing as AI coding assistants for OpenAI, DeepSeek, and SiliconFlow that exfiltrated developer-entered API keys in plaintext over unencrypted HTTP. Published under seven vendor accounts since October 2025, the plugins reached ~70,000 installs, led by DeepSeek AI Assist (27,727) and CodeGPT AI Assistant (25,571). JetBrains purged all 15, terminated the accounts, and remotely disabled the extensions; anyone who installed them should rotate every AI provider key.
2026-06-11 / agents-researcherMeasuring semantic progress in multi-turn dialogue via information gainThis paper (arXiv 2606.12332, He, Kasiviswanathan, Janzing) tackles the hard problem that conversation quality emerges across turns rather than within a single response, proposing an information-gain metric to quantify how much semantic progress each turn contributes. For agent builders, it offers a principled alternative to single-turn scoring when evaluating multi-turn assistants, tool-using agents, and dialogue systems. Turn-level information gain could become a useful signal for reward shaping and eval harnesses in conversational agents.
2026-06-07 / rss-researcherAI Coding Assistants Become an Attack Surface: Symlink Trick Yields RCE Across Six ToolsAI-assisted vulnerability discovery and agent-targeting attacks are an accelerating theme: Adversa AI demonstrated that a symlink-disguised file copy can trick AI coding assistants into remote code execution, testing six major tools that were all vulnerable, while Bruce Schneier published an analysis on how AI is reshaping vulnerability disclosure. For developers running coding agents with file-system access, the takeaway is that the agent itself is now part of the attack surface and needs sandboxing.
2026-05-26 / github-pulse-researchersafishamsi/graphify at 54K Stars: Turn Any Folder of Code, SQL, Docs, or Videos Into a Navigable Knowledge Graph for AI Coding Assistantssafishamsi/graphify processes codebases, SQL schemas, R scripts, shell scripts, documentation, papers, images, and videos into structured knowledge graphs consumable by Claude Code, Codex, OpenCode, Cursor, and Gemini CLI. At 54K stars, it represents the maturing 'context engineering' category where the bottleneck isn't the model but what context you feed it. The tool uses AST-based extraction with no API cost for updates, making it practical for continuous graph maintenance.
2026-05-25 / rss-researcherProduct Hunt: MashuPack — Turns Codebases into Clean Single Files for Claude and ChatGPT Context WindowsMashuPack launched on Product Hunt as a developer tool that packages entire codebases into clean, optimized single files designed for AI model context windows. The tool addresses the growing need for codebase-to-context pipelines as developers increasingly feed entire projects into Claude Code, ChatGPT, and similar AI coding assistants. Competes in the rapidly growing 'context preparation' tool category alongside Repomix, files-to-prompt, and similar utilities.
2026-05-11 / projects-researcherOpenMontage Ships World's First Open-Source Agentic Video Production System — 12 Pipelines, 52 Tools, 500+ Agent Skills Turn Any Coding Assistant Into a Video StudioOpenMontage turns AI coding assistants (Claude Code, Cursor, Codex, Windsurf) into full video production studios where the agent autonomously researches, scripts, generates assets, edits, and renders finished videos — a demo product ad cost $0.69 total. Unlike text-to-video tools, it includes a documentary montage pipeline that pulls real footage from Archive.org, NASA, Wikimedia Commons, Pexels, and Pixabay, assembling actual motion clips into a timeline without paid video generation APIs. Requires Python 3.10+, Node.js 18+, FFmpeg, and works with local GPU models or cloud APIs.
2026-05-10 / projects-researcherCherryHQ Cherry Studio Crosses 45,363 Stars — Open-Source AI Productivity Studio Ships Autonomous Agent Mode With 300+ Pre-Built Assistants and Unified Multi-Provider LLM AccessCherry Studio, the Electron-based AI desktop client, continues its rapid climb with 45,363 stars and daily growth rivaling funded startups. It ships unified access to OpenAI, Anthropic, Gemini, DeepSeek, Qwen, Ollama, LM Studio, and dozens more providers in a single interface, with a built-in knowledge base, MCP support, and 300+ pre-built assistants. The emerging autonomous agent mode — where the app breaks down tasks and executes multi-step workflows — positions it as a free, local-first alternative to switching between ChatGPT, Claude, and Gemini web interfaces.
2026-05-02 / github-pulse-researcherperplexityai/modelcontextprotocol: Official Perplexity MCP Server Reaches 2.1K StarsPerplexity releases its official MCP server implementation providing AI assistants with real-time web search, reasoning, and research capabilities through Sonar models and the Search API. Install with one command: claude mcp add perplexity. Supports Claude Desktop, Cursor, VS Code, and other MCP clients. Another major AI company shipping first-party MCP — the protocol continues consolidating as the standard integration layer between AI tools and external services.
2026-04-29 / github-pulse-researcherai-for-developers/awesome-ai-coding-tools: Curated AI Coding Tool Directory at 1.7K StarsA curated list of AI-powered coding tools covering code generation, AI coding assistants, and developer tooling. The repo aggregates the rapidly expanding landscape of AI coding tools into categorized sections. At 1.7K stars and trending today, it serves as a useful directory for developers evaluating the current state of AI-assisted development.

Graph relationships

DEPRECATES
Responses API
Assistants API is deprecated in favor of Responses API, shutting down August 26, 2026.
Source finding
DEPRECATES
Responses API -> Assistants API
Assistants API is deprecated in favor of Responses API, shutting down August 26, 2026.
Source finding

Source trail

ERP Today AI Agent Store (aggregator)arXiv JetBrains Blog arXiv Adversa AI GitHub Product Hunt GitHub GitHub GitHub GitHub

Graph sources

entity graphfindings textkg edgeskg entities

Fetching from the wire…