TIL: Claude Code local sandbox environment is open-source. > native OS sandboxing primitives (sandbox-exec on macOS, bubblewrap on Linux) and proxy-based network filtering. It can be used to sandbox the behaviour of agents, local MCP servers, bash commands and arbitrary processes.

Yesterday, we shipped the File Search Tool, a fully managed RAG system integrated into the Gemini API. Excited to share the everything you need to know as a Web (JavaScript) Developer, covering creating and finding file search stores, concurrent and advanced uploads techniques, and how to update, find, or delete documents. - Create persistent File Search Stores and find them by display name. - Upload multiple files concurrently with custom chunking and metadata. - Run standard RAG generation queries and find specific indexed documents. - Update, delete, and cleanup documents and File Search Stores. Link to guide below :)

We added the missing code snippets for JavaScript in the Gemini API File Search documentation:

We've launched the File Search Tool, a fully managed RAG system integrated into the Gemini API that simplifies grounding models with your private data to deliver more accurate, verifiable responses. - $0.15/m tokens for indexing, free storage and embedding generation at query time. - Free tier storage 1 GB, tier 1: 10 GB: Tier 2: 100 GB; Tier 3: 1 TB - Vector search powered by the Gemini Embedding model. - Supports PDF, DOCX, TXT, JSON, and programming language files. - Includes citations specifying exact source documents used. - Combines results from parallel queries in under 2 seconds.

Here is an example of using a deeply nested Pydantic Schema with +30 fields, Union, Optionals and recursive structures with the new Sturctured Structured Outputs in the Gemini API! 💥 Note: To enable this you need to use the new `response_json_schema` parameter instead of the old `response_schema`. Documentation and Snippet below.

Enhancements to Structured Outputs in the Gemini API, expanding JSON Schema support for more predictable results, Unions and Recusrive Schema. - Supports `anyOf` for unions and `$ref` for recursive schemas. - Compatible out-of-the-box with Pydantic and Zod libraries. - Preserves output property ordering to match input schemas. - Includes minimum and maximum, prefixItems and more. - Streams partial JSON chunks for faster processing. We have multiple examples for Unions or recursive schema as part of our new documentation. 📕

I really like clean visuals. 🥵

QoL Update! 🔥 You can now prefill prompts via URL parameter (`prompt=`) in @GoogleAIStudio for Gemini! Append `?prompt=Your Text` to ai(.)studio/prompts/new_chat Example → ai(.)studio/prompts/new_chat?prompt=Write a Haiku about flowers Boom ready to run! 💥

Introducing the Gemini Docs MCP Server, a local STDIO server for searching and retrieving Google Gemini API documentation. This should help you build with latest SDKs and model versions. 🚀 - Run the server directly via uvx without explicit installation. - Performs full-text search across all Gemini documentation pages locally. - Passed 114/117 for Python and Typescript using latest SDKs and Model. - 3 Tools: search_documentation, get_capability_page, get_current_model. - Utilizes a local SQLite database with FTS5 for efficient querying. - Works with @code, @cursor_ai, Gemini CLI … and every other tool support MCP.

Last week in AI: - OpenAI restructured into a Public Benefit Corp with a $1.4T compute roadmap. - Poolside raised 1B at 12B valuation and Eric Zelikman secured $1B for a new venture. - Gemini API prices slashed: 50% off Batch API and 90% off context caching. - Google AI Pro plans (with Gemini 2.5) are rolling out via Jio in India. - Moonshot AI shipped Kimi Linear, a hybrid model with 6x decoding throughput. - Hugging Face released the 200+ page "Smol Training Playbook" for LLMs. - vLLM introduced "Sleep Mode" for instant, zero-reload model switching. - New findings suggest switching from BF16 to FP16 reduces RL fine-tuning divergence. - Epoch AI analysis shows open-weight models now catch up to closed SOTA in ~3.5 months.