Cycle on Jun 8, 2026

I will start by listing the topic's open threads to see what specific questions or watches are currently registered for this research topic.

I will read the existing notes for the topic to see how the market map is currently defined and what findings we have on Gemini Deep Research, Perplexity Deep Research, Sacra, and academic AI tools.

Let's review the notes we have. We have notes on: - `market-map-positioning-hey-lefty` - `academic-ai-research-scholar-vs-synthesis` - `openai-deep-research-api-and-pro-tier` - `perplexity-deep-research-consumer-agent` - `gemini-deep-research-agent` - `sacra-premium-private-market-mcp-data` - `integrating-custom-mcp-servers` Wait, since it's June 2026, let's do some research on Sacra, Gemini Deep Research, Perplexity Deep Research, Google Scholar, and see if there are new updates, or if we can refine the market map. Let's first search for Sacra's current pricing, business model, and features. Let's do a search for "Sacra research platform" or "Sacra pricing" or "Sacra API". Let's also do a search on other competitors in the market for autonomous or AI research tools. Let's see what's happening. Let's search the web for Sacra.

Let's fetch Sacra's pricing page and documentation to understand their pricing tiers, business model, and API features. Let's fetch `https://sacra.com/pricing/` and `https://docs.sacra.com/`. Let's also fetch `https://sacra.com/p/introducing-the-sacra-revenue-api/`.

The Sacra pricing page shows: - Standard: $50/mo (1 seat, 500 tasks, company & market reports, charts & datasets, expert interviews, Model Context Protocol (MCP) server access, API, revenue & valuation models, XLS & PDF export, bespoke research) - Pro: $350/mo (5 seats, 3500 tasks, same features) - Platform: $1,500/mo (20 seats, 25,000 tasks, same features) - Enterprise: Custom Wait, let's fetch `https://docs.sacra.com/` to see the API capabilities and what datasets they provide. Let's do that.

Ah, let's search for "Sacra MCP server" or "Sacra Model Context Protocol" to see how they integrate with agents. This is highly relevant for our market map and positioning. Let's do a web search.

Let's fetch `https://sacra.com/features/mcp/` and `https://sacra.com/p/introducing-sacra-connector-for-claude/` (or similar blog post) to see how Sacra's MCP server works and how it represents private market intelligence as agent-ready infrastructure. Let's do that.

Wait, let's fetch `https://sacra.com/p/introducing-sacra-connector-for-claude/` or similar. Let's do a quick search for the exact URL or find a LinkedIn post or other info about the Sacra MCP connector. Let's do a search for "Introducing Sacra MCP connector for Claude" or "Unbundle Sacra reports with Claude".

This is brilliant! Let's fetch `https://sacra.com/p/introducing-sacra-connector-for-claude/`, `https://sacra.com/p/introducing-sacra-mcp-connector-chatgpt/`, `https://sacra.com/p/build-on-sacra-mcp/`, and `https://sacra.com/p/sacra-for-slack/`. These will give us deep insight into how Sacra is leveraging MCP and agent-ready infrastructure, which is a key part of our market map and how Hey, Lefty fits. Let's fetch them one by one. Let's start with `https://sacra.com/p/introducing-sacra-connector-for-claude/` and `https://sacra.com/p/build-on-sacra-mcp/`.

Let's fetch `https://sacra.com/p/build-on-sacra-mcp/` to see how other companies are building on top of it. Let's do that.

This is incredibly relevant! Sacra has unbundled itself by providing a server-side MCP server at `https://mcp.sacra.com/mcp` which can be integrated into custom agent workflows (like Goodfin Go, a React-based agent). Let's see: what are the other platforms mentioned in the brief? - Gemini Deep Research (we have a note on `gemini-deep-research-agent`) - Google Scholar (we have a note on `academic-ai-research-scholar-vs-synthesis`) - Perplexity (we have a note on `perplexity-deep-research-consumer-agent`) - Sacra (we have a note on `sacra-premium-private-market-mcp-data`) Wait, let's look at the note `market-map-positioning-hey-lefty` to see how we can refine the market map. Let's fetch the contents of `market-map-positioning-hey-lefty` first, or we can read it. Oh, we don't have a direct tool to read a specific note, but we can update it if we write to its slug. Wait, can we read its content? Let's check `read_notes` again. Ah, `read_notes` only returned previews. Let's see if we can search for the note or if we can fetch it, or if we should just write a comprehensive update. Wait, we can find out if there are other ways to read notes. No, `read_notes` returns the list of notes with previews. Wait, let's search for "Gemini Deep Research" and "Perplexity Deep Research" and see what's changed in 2026. Let's do a web search on "Gemini Deep Research" to see if there's any news, pricing updates, or API changes since April 2026. Let's do that.

Let's fetch `https://blog.google/innovation-and-ai/models-and-research/gemini-models/next-generation-gemini-deep-research/` or `https://ai.google.dev/gemini-api/docs/interactions/deep-research` or `https://vmblog.com/news/lovelace-matches-gemini-deep-research-at-less-than-1-of-the-cost/` to see what is happening with Gemini Deep Research Max and how it compares to Lovelace or others. Let's start with `https://ai.google.dev/gemini-api/docs/interactions/deep-research` to get the official Google developer API details.

This is fantastic! Google's Gemini Deep Research Agent has: - Two versions: `deep-research-preview-04-2026` (speed & efficiency) and `deep-research-max-preview-04-2026` (maximum comprehensiveness). - Features: Collaborative planning (multi-turn planning where the agent proposes a research plan and the user can review/modify/approve before execution), Visualizations (generating charts, graphs, and visual elements dynamically), and Supported tools (Google Search, URL Context, Code Execution, MCP Server, File Search). - Supported tools now natively includes Model Context Protocol (MCP Server)! This is a critical development since our last cycle. Gemini Deep Research Agent now supports connecting to remote MCP servers for external tool access. Let's see if we can find more about this MCP integration. Let's do a search on Google's Gemini API MCP server support.

Let's fetch `https://blog.google/innovation-and-ai/models-and-research/gemini-models/next-generation-gemini-deep-research/` or `https://aistudio.google.com/learn/deep-research-developer-guide` or `https://www.philschmid.de/deep-research-update` to see the details of Google's next-generation Gemini Deep Research announcement. Let's search for "Deep Research Max: a step change for autonomous research agents" or fetch the Phil Schmid blog post. Let's do a search for the exact text of Google's announcement.

Let's fetch `https://blog.google/innovation-and-ai/models-and-research/gemini-models/next-generation-gemini-deep-research/`.

This is incredibly powerful! Let's examine the details: - Released on April 21, 2026, by Lukas Haas and Srinivas Tadepalli (Product/Program Managers at Google DeepMind). - Built with Gemini 3.1 Pro. - Two agents: - **Deep Research**: Optimized for speed & efficiency, replacing the December preview, lower latency/cost. - **Deep Research Max**: Optimized for maximum comprehensiveness, leveraging extended test-time compute to iteratively reason, search, and refine. - New features: - **Model Context Protocol (MCP) support**: Securely connect Deep Research to custom private data or specialized professional streams (e.g., financial/market data providers). - **Collaborative planning**: Review, guide, and refine the research plan generated by the agent before execution. - **Native charts and infographics**: Generates high-quality charts/infographics in-line with HTML or "Nano Banana" (which seems to be a Google visualization technology). - **Extended tooling**: Google Search, remote MCP servers, URL Context, Code Execution, File Search simultaneously, or turn off web access entirely to search custom data only. - **Multimodal grounding**: Inputs of PDFs, CSVs, images, audio, video. - **Real-time streaming**: Thought summaries, text/image outputs as they are generated. - Collaborations: Working with **FactSet**, **S&P Global**, and **PitchBook** on their MCP server designs to let shared customers integrate financial data offerings into workflows powered by Deep Research. Wait, let's look at our watches. One of our watches is: - `019ea05f-eaaf-7afb-bb51-55b5f00e17d3`: "Institutional Research Platforms Launch Public MCP Servers" - Description: "Monitor when major financial or market data platforms (e.g., PitchBook, S&P Global, FactSet) release developer-facing or public MCP servers." - Trigger: "Fires when a major institutional competitor to Sacra or developer data provider launches a public or partner-accessible MCP server." - Wait, does this April 21, 2026 Google announcement mean they are collaborating with FactSet, S&P Global, and PitchBook on their MCP server designs? Let's check if there are any actual public or partner-accessible launches of these MCP servers. Let's do a web search. Let's search for "FactSet MCP server", "S&P Global MCP server", "PitchBook MCP server".

Wow! This is a massive development! Major financial and market data platforms (FactSet, S&P Global, PitchBook) have indeed released developer-facing/public/partner-accessible MCP servers! Let's see: - FactSet released a Model Context Protocol (MCP) server providing real-time, production-grade access to trusted financial intelligence for AI workflows. The press release title is "FactSet Meets Demand for AI-Ready Data, First to Announce MCP...". - S&P Global Kensho LLM-ready API provides an MCP server enabling Claude users to connect directly. - PitchBook Premium connects via MCP. Let's fetch `https://developer.factset.com/mcp` and `https://finance.yahoo.com/news/factset-meets-demand-ai-ready-110500592.html` or similar. Let's do a search to get the details of FactSet's MCP server launch. Let's fetch `https://developer.factset.com/mcp` or `https://insight.factset.com/enterprise-mcp-model-context-protocol-part-one`. Wait, let's search for "FactSet Meets Demand for AI-Ready Data, First to Announce MCP" to get the details of the launch and see if it fires our watch! This is a dated catalyst or ongoing watch that has fired! Let's search and fetch.

Let's fetch `https://investor.factset.com/news-releases/news-release-details/factset-meets-demand-ai-ready-data-first-announce-mcp-sans` or `https://www.stocktitan.net/news/FDS/fact-set-meets-demand-for-ai-ready-data-first-to-announce-mcp-sans-cg0hxqvudtyk.html` to get the details of the FactSet press release. This is very important. Let's fetch the stocktitan one or the investor.factset.com one. Let's do `https://www.stocktitan.net/news/FDS/fact-set-meets-demand-for-ai-ready-data-first-to-announce-mcp-sans-cg0hxqvudtyk.html`.

This is an extremely important piece of news! Let's examine this carefully. - FactSet launched its Model Context Protocol (MCP) server on **December 16, 2025**. It is the first production-grade MCP server in the financial intelligence industry, providing direct, real-time access to 9 key datasets (Fundamentals, Consensus Estimates, Ownership, Global M&A, Global Pricing, People, Events, Supply Chain, and Geographic Revenue Exposure) without intermediaries. - The launch followed a beta with **45 firms** and over **800 institutional users**. - This is a direct trigger for our watch `019ea05f-eaaf-7afb-bb51-55b5f00e17d3`: "Institutional Research Platforms Launch Public MCP Servers". S&P Global and PitchBook have also launched MCP servers or integrations! Let's see if we can do a search on S&P Global's MCP server and PitchBook's MCP server to get their details as well, because this completes the picture of the "institutional tier" adopting MCP. Let's do a search.

Let's fetch `https://docs.kensho.com/llmreadyapi/mcp` and `https://www.spglobal.com/en/press/press-release/sp-global-delivers-trusted-financial-data-and-insights-to-customers-through-app-for-chatgpt` or `https://www.prnewswire.com/news-releases/sp-global-and-anthropic-announce-integration-of-sp-globals-trusted-financial-data-into-claude-302505482.html` to see the details of S&P Global's MCP launch. This is amazing. Let's fetch `https://www.spglobal.com/en/press/press-release/sp-global-delivers-trusted-financial-data-and-insights-to-customers-through-app-for-chatgpt`.

Ah, that URL failed to fetch. Let's try to fetch another URL: `https://www.prnewswire.com/news-releases/sp-global-and-anthropic-announce-integration-of-sp-globals-trusted-financial-data-into-claude-302505482.html` or `https://docs.kensho.com/llmreadyapi/mcp`. Let's search for "S&P Global and Anthropic Announce Integration of S&P Global's Trusted Financial Data into Claude" or search for "Kensho LLM-ready API MCP server". Let's do a web search.

Let's fetch `https://docs.kensho.com/llmreadyapi/overview` to see how S&P Global / Kensho are positioning their MCP server and what data they offer. Let's do that.

Ah, let's fetch `https://docs.kensho.com/llmreadyapi/mcp` to see the details of the MCP server. Let's do that.

Ah, the web search budget for this cycle is spent. This is a very clear signal to wrap up our research, write our notes, and finish the cycle. Let's carefully analyze what we have. First, let's look at our threads. We have three open threads: 1. `019ea05f-eaaf-7afb-bb51-55b5f00e17d3` (ongoing): "Institutional Research Platforms Launch Public MCP Servers" - Trigger: "Fires when a major institutional competitor to Sacra or developer data provider launches a public or partner-accessible MCP server." - FactSet launched its production-grade MCP server on **December 16, 2025**, giving direct, real-time access to Fundamentals, Consensus Estimates, Ownership, Global M&A, Global Pricing, People, Events, Supply Chain, and Geographic Revenue Exposure. - S&P Global (via Kensho) launched its LLM-Ready API MCP Server, connecting S&P Capital IQ Financials, Market Data, Business Relationships, Earnings Call Transcripts, Company Intelligence, M&A Transactions directly to Claude, ChatGPT, Microsoft Copilot Studio, Amazon Quick, Databricks, and Mistral. - Google DeepMind's Deep Research Max announced in April 2026 that it is collaborating with FactSet, S&P Global, and PitchBook on their MCP server designs to let shared customers integrate financial data into Deep Research workflows. - This means this watch has DEFINITELY fired! Since its cardinality is `ongoing`, we should record an event on it with `resolves: false`. Let's check the other watches: 2. `019ea05f-df8d-7be9-9dd9-99ddfd8136bb` (one_shot): "OpenAI Releases o4 Deep Research API Model" - Did this happen? We don't have concrete news of an "o4-deep-research" API model launch in this cycle's evidence. We should keep this open. 3. `019ea05f-d685-7208-a431-55568fba4987` (one_shot): "Gemini Deep Research Max Transitions to GA" - Did this happen? In Google's April 21, 2026 blog post, they announced that Deep Research and Deep Research Max are available "in public preview via paid tiers in the Gemini API", and "will also soon be available to startups and enterprises in Google Cloud." So they are still in public preview. This watch should remain open. Let's record the event on `019ea05f-eaaf-7afb-bb51-55b5f00e17d3` (the ongoing institutional MCP server watch), and then write our notes. Let's call `record_event` for the institutional MCP server watch. - `thread_id`: `019ea05f-eaaf-7afb-bb51-55b5f00e17d3` - `cycle_id`: `6f68fc89-298c-4d65-adb6-83709c77f39e` (Wait, let's get the cycle_id from the user prompt or leave it empty? The user prompt has "cycle-6f68fc89-298c-4d65-adb6-83709c77f39e". Let's pass `6f68fc89-298c-4d65-adb6-83709c77f39e` as the cycle_id.) - `summary`: "FactSet and S&P Global launch production-grade Model Context Protocol (MCP) servers for direct enterprise AI access." - `outcome`: "Fired. FactSet released its MCP server in Dec 2025; S&P Global (Kensho) offers remote and local MCP servers for Capital IQ data." - `resolves`: `false` (ongoing watch) - `salience`: 5 (very high) Let's call `record_event` now.

Now let's synthesize our findings and write/update notes. Let's see: we have several specific topics: 1. **Sacra**: We need to update our note on Sacra: `sacra-premium-private-market-mcp-data`. Let's see what we learned about Sacra: - Pricing: Standard is $50/mo (1 seat, 500 tasks, company & market reports, charts, expert interviews, MCP server access, API). Pro is $350/mo (5 seats, 3500 tasks). Platform is $1,500/mo (20 seats, 25,000 tasks). - Features & API: Includes company profiles, funding rounds, news, and document contents. - MCP Server: Sacra provides a server-side MCP server at `https://mcp.sacra.com/mcp` which can be integrated into custom agent workflows (like Goodfin Go, a React-based agent). - Value Prop: Unbundling private market intelligence as "agent-ready infrastructure", protecting AI workflows from SEO spam and web trash. 2. **Gemini Deep Research**: We need to update our note on Gemini: `gemini-deep-research-agent`. - Released on April 21, 2026, by Lukas Haas and Srinivas Tadepalli (Google DeepMind). - Built on Gemini 3.1 Pro. - Two versions: `deep-research-preview-04-2026` (speed & efficiency, low latency) and `deep-research-max-preview-04-2026` (maximum comprehensiveness, extended test-time compute). - Key Features: - Model Context Protocol (MCP) support: Securely connect to custom data and specialized professional data streams. - Collaborative planning: Review, guide, and refine the research plan before execution. - Native charts and infographics: Generates high-quality charts/infographics in-line with HTML or "Nano Banana" (Google's visualization tech). - Extended tooling: Google Search, remote MCP, URL Context, Code Execution, File Search simultaneously. - Multimodal grounding: Support PDFs, CSVs, images, audio, video. - Real-time streaming: Thought summaries and text/image outputs. - Collaborations: Working with FactSet, S&P Global, and PitchBook on their MCP server designs to let shared customers integrate financial data into Deep Research workflows. 3. **Institutional MCP Servers (FactSet & S&P Global)**: This is a massive development that warrants a new note or can be integrated into the market map and/or a new note. Wait, let's look at the market map note: `market-map-positioning-hey-lefty`. Let's write a dedicated note on the institutional MCP server trend (e.g. `institutional-mcp-financial-data-servers`) and update the `market-map-positioning-hey-lefty` note to reflect this shift. Wait, let's look at what the market map currently represents. The previous findings mention "two distinct tiers in 2026" or "three distinct tiers". Let's see: - Tier 1: Generalist high-speed search/synthesizers (Perplexity Deep Research, OpenAI Deep Research). - Tier 2: Specialized research platforms providing "agent-ready infrastructure" (Sacra for private markets, FactSet/S&P Global/PitchBook for institutional public/private markets via MCP servers). - Tier 3: Model-agnostic multi-source orchestrators (like Hey, Lefty) that tie these tiers together. Wait! This is an incredibly elegant market map! Let's design the notes: - Update `sacra-premium-private-market-mcp-data` with their new pricing tiers ($50/mo Standard to $1,500/mo Platform) and the details of their server-side MCP (`mcp.sacra.com/mcp`) and how Goodfin built on top of it. - Update `gemini-deep-research-agent` with the April 2026 release of Gemini 3.1 Pro-powered Deep Research and Deep Research Max, detailing collaborative planning, native visualizations, MCP support, and collaborations with PitchBook, FactSet, and S&P Global. - Create a new note `institutional-mcp-financial-data-servers` detailing FactSet's "MCP Sans Intermediary" (released Dec 16, 2025) and S&P Global's Kensho LLM-Ready API MCP Server, showing how institutional financial intelligence is being unbundled into standardized, agent-ready APIs. - Update `market-map-positioning-hey-lefty` to synthesize this entire 2026 landscape: how the unbundling of premium private market (Sacra) and institutional public market (FactSet, S&P Global, PitchBook) data via MCP servers creates the ultimate opportunity for **Hey, Lefty** to position itself as the supreme model-agnostic multi-source orchestrator. Hey, Lefty doesn't need to build its own financial or private company databases; it can orchestrate these powerful MCP servers using advanced autonomous agents (like Gemini Deep Research Max or custom orchestrators) to deliver unified, expert-grade synthesis. Let's write the notes now. Let's start with `sacra-premium-private-market-mcp-data`. Let's draft the markdown for this note first. Wait, we must follow the rules: - Carry 1-2 REPRESENTATIVE VERBATIM QUOTES, copied exactly, each kept with the link it came from. - Link other findings using wikilinks like `[[its-slug]]`. - No YAML frontmatter. The title is the first `# ` heading and the slug is the slug parameter we pass. Let's write `sacra-premium-private-market-mcp-data`.

Now let's write or update `gemini-deep-research-agent` with the April 2026 release of Gemini 3.1 Pro-powered Deep Research and Deep Research Max. Let's draft the markdown for this note. Let's do that.

Now let's write a new finding note specifically on the emergence of institutional financial MCP servers, titled `institutional-mcp-financial-data-servers`. This is a critical development. S&P Global and FactSet have unbundled their data. Let's write this note.

Now, let's update the main market map and positioning note `market-map-positioning-hey-lefty` to incorporate these incredible developments. Let's design the updated market map structure: - **Tier 1: Generalist High-Speed Search & Synthesis Engines** (e.g., Perplexity Deep Research, OpenAI Deep Research, Gemini Deep Research). Focus: Web-scale synthesis, high speed, consumer-friendly subscriptions or developer-centric raw APIs. - **Tier 2: Specialized Data Providers as Agent-Ready Infrastructure** (e.g., Sacra for pre-IPO/private markets, FactSet & S&P Global via MCP for institutional public markets). Focus: Exposing curated, high-fidelity databases directly to LLMs via standardized MCP protocols, eliminating "web junk" and SEO spam in professional workflows. - **Tier 3: Model-Agnostic Multi-Source Orchestrators (Hey, Lefty)**. Focus: Orchestrating across both Tier 1 reasoning engines and Tier 2 high-fidelity data providers. By leveraging custom MCP servers (like [[integrating-custom-mcp-servers]]), Hey, Lefty acts as the unified control plane. It coordinates multi-step, multi-source research cycles, running specialized agents to query Sacra, FactSet, or academic databases simultaneously, then synthesizes the results into structured, expert-grade notes. Let's write this updated note. This is a beautiful, cohesive synthesis of the entire market landscape in 2026! Let's do it.

Excellent. All of our notes have been successfully written and updated: - `sacra-premium-private-market-mcp-data` (Updated) - `gemini-deep-research-agent` (Updated) - `institutional-mcp-financial-data-servers` (Created) - `market-map-positioning-hey-lefty` (Updated) Let's check if we need to resolve any threads. We have three threads listed: - `019ea05f-eaaf-7afb-bb51-55b5f00e17d3` (watch): "Institutional Research Platforms Launch Public MCP Servers" - Wait, this is a watch thread. Can we resolve it? No, it's an ongoing watch. We already recorded an event on it (using `resolves: false` as it is ongoing). - `019ea05f-df8d-7be9-9dd9-99ddfd8136bb` (watch): "OpenAI Releases o4 Deep Research API Model" - This is a watch thread, one_shot. It hasn't fired yet, so we leave it open. - `019ea05f-d685-7208-a431-55568fba4987` (watch): "Gemini Deep Research Max Transitions to GA" - This is a watch thread, one_shot. It is still in preview, so we leave it open. Wait, are there any other open threads that we addressed? No, `list_threads` only showed these three watch threads. We don't have any standard question/investigation threads to resolve this cycle. Let's call `finish_cycle` to complete this research cycle. - `no_findings_reason`: (Omit or empty string since we wrote findings) Let's call `finish_cycle`.

I have completed this research cycle for **Hey, Lefty's** competitive positioning and market mapping. Here is a summary of the critical updates and findings established in this cycle: ### 1. The 2026 Competitive Landscape: A Three-Tier Market Map The market for autonomous AI research tools has stratified into three distinct tiers. We have updated our core positioning framework to reflect this maturity: * **Tier 1 (Generalist Reasoning Engines):** Includes **Gemini Deep Research Max** (powered by Gemini 3.1 Pro, released in April 2026), **OpenAI Deep Research**, and **Perplexity Deep Research**. These tools excel at web-scale, long-horizon synthesis using extended test-time compute, but are primarily generalist engines heavily reliant on public web search. * **Tier 2 (Specialized Data Providers as Agent-Ready Infrastructure):** Premium vertical platforms that have unbundled their databases into standardized **Model Context Protocol (MCP)** endpoints. * **Sacra** exposes private company fundamentals, funding rounds, and analyst transcripts via `mcp.sacra.com/mcp` to protect AI agents from public-web SEO spam. * **FactSet** launched its production-grade "MCP Sans Intermediary" server in December 2025, exposing nine core financial datasets. * **S&P Global (Kensho)** offers a hosted and local MCP server connecting S&P Capital IQ financials, transcripts, and M&A data directly to Claude, ChatGPT, and enterprise data platforms. * **Tier 3 (Model-Agnostic Multi-Source Orchestrators):** This is where **Hey, Lefty** fits. By acting as a unified control plane, Hey, Lefty coordinates across Tier 1 reasoning models and Tier 2 high-fidelity data providers via custom MCP servers, compiling findings into structured, persistent knowledge blocks (notes) rather than one-off chat summaries. ### 2. Key Findings & Persistent Notes Updated * **[[market-map-positioning-hey-lefty]] (Updated):** Synthesized the complete 2026 three-tier market map. Established Hey, Lefty's strategic positioning: avoiding the capital-intensive task of rebuilding proprietary databases, standardizing on MCP orchestration to let users connect their own FactSet/Sacra subscriptions, and focusing on multi-model synthesis and structured, persistent memory. * **[[gemini-deep-research-agent]] (Updated):** Documented Google DeepMind's April 21, 2026 release of **Deep Research** and **Deep Research Max** (Gemini 3.1 Pro). Key features analyzed include **collaborative planning** (user-guided research plans), **native charts & infographics**, and their active collaborations with FactSet, S&P Global, and PitchBook on MCP server designs. * **[[sacra-premium-private-market-mcp-data]] (Updated):** Outlined Sacra's new tiered pricing model (ranging from the $50/mo Standard plan to the $1,500/mo Platform plan) and detailed how companies like **Goodfin** are programmatically calling Sacra’s MCP server (`mcp.sacra.com/mcp`) from their backend agent architectures to bypass public web noise. * **[[institutional-mcp-financial-data-servers]] (New Note Created):** Detailed the massive structural shift where **FactSet** and **S&P Global** have launched production-grade, developer-facing MCP servers, representing the unbundling of institutional financial intelligence for direct, real-time AI consumption. ### 3. Thread & Watch Updates * **Fired Watch Registered:** We recorded a high-salience event on our ongoing watch thread (*"Institutional Research Platforms Launch Public MCP Servers"*), confirming that FactSet and S&P Global have successfully launched production-grade MCP servers and are actively collaborating with Google DeepMind on Deep Research integrations. * **Ongoing Watches Maintained:** We continue to monitor OpenAI's potential developer release of an `o4-deep-research` API model and the transition of Gemini Deep Research Max from public preview to general availability (GA) in the Gemini API.