> For the complete documentation index, see [llms.txt](https://knowledgebase.fabricdata.com/insights/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://knowledgebase.fabricdata.com/insights/releases/mcp.md). # MCP ## Releases Deployment history for the **Origin Insights MCP Server**. You don't need to update anything on the client side: when you connect, you always have the latest version. {% updates format="full" %} {% update date="2026-08-05" %} ## v.1.3.2 **Know your audience.** This release introduces a rebuilt audience intelligence report, faster and more reliable documentation answers, and a set of accuracy and stability improvements across reports. No breaking changes. #### New capabilities * **Audience report, rebuilt** — ask who watches a title and get a decision-ready profile: interactive demographics (filter the whole panel by gender with one click), audience segment affinity, and advertising-ready interest categories (IAB). Profiles are served from a curated audience dataset for faster, more consistent answers — with on-demand analysis as a fallback — and the layout now reads well on a phone. #### Quality & reliability * **Better documentation answers** — the knowledge base behind "ask the docs" is now served directly by the MCP, understands questions asked in Spanish against the English documentation, and distinguishes a real outage from a simple "no match." * **Numbers you can trust** — demand scores in title reports are shown exactly as computed (no rounding drift), and every report's *Key Findings* are validated against what the report actually displays. * **Search arguments honored** — availability queries now respect the title and country arguments exactly as advertised. * **Sturdier deployments** — releases are pinned and roll back automatically if a deployment misbehaves, reducing the chance of a bad rollout ever reaching you. #### Compatibility No breaking changes and nothing to update on your side. No tools were removed or renamed. {% endupdate %} {% update date="2026-07-22" %} ## v.1.3.1 Release of new capabilities and hardening on top of an unchanged production configuration. Adds decision-ready composite analyses, data and deck export, a better similarity engine, and several security and reliability improvements. No breaking changes. #### New capabilities * **Market Radar (new report)** — summarizes what moved in a market over a recent window versus the equivalent previous window: who is rising and falling in demand (week over week), catalog entries and exits, and rental/purchase price changes. Can be scoped by country and platform. * **4 analyst "playbook briefs"** — composite analyses that synthesize a decision, not just a list: what to acquire for a region (ranked shortlist of titles with demand but no local availability), whether to renew a license in a market (Renew / Renegotiate / Drop verdict with evidence), which owned titles are underexploited (demand vs. distribution), and which genres to invest in for a region (per-genre greenlight brief). * **Data export** — download the full catalog to CSV (via presigned link, no practical limit) or to Excel (.xlsx, capped), export the result of a custom query to CSV, and a new `export` argument on \~12 listing tools to download the same filtered query in one step. In addition, every report now offers export to multi-sheet Excel, PPTX deck, Brandbook PDF deck, and self-contained HTML. * **Embedding-based title similarity** — `find_similar_titles` now uses an embeddings engine as the primary path (with automatic fallback to the genre + keywords method if the service does not respond), returning more relevant neighbors ranked by similarity and demand. * **Prompt gallery in the MCP** — 7 starter prompts visible in the connector's "+" menu for onboarding, plus a new weekly market movement summary resource. #### Quality and reliability * **Richer, more data-honest reports** — dedicated report apps for box office and people, aggregates computed over the full dataset when the detail is truncated (totals/averages no longer come only from the visible rows), and a per-dataset stale-data warning that shows when a report may be serving old information instead of presenting it as current. * **Fewer unnecessary reports** — a density heuristic decides when to generate a report based on data volume and shape, and the report summary (headline, executive bullets, and metrics) is passed inline to the assistant so it can narrate without a second call. * **More stable under error** — malformed user SQL errors no longer count against the circuit breaker (avoids spurious trips), and there is per-tool cost telemetry (bytes scanned and latency). * **Security — 5 fixes** — per-browser CSRF binding on OAuth login, client IP correctly taken from `X-Forwarded-For` (rate-limit anti-spoofing), image proxy hardening (rate limit + byte cap), access control on `/metrics` (the identity/cost list is redacted for non-privileged callers), and closure of a SQL row-cap bypass. #### Compatibility No breaking changes and nothing to update on your side. No tools were removed or renamed; existing response schemas only add new fields. The active production surface goes from 102 to 110 tools (+8), and the MCP protocol adds the `list_prompts` / `get_prompt` methods. **Count note:** starting with v1.3.1, the inventory counts only active capability tools accessible in production. Recalculating v1.3.0 with that criterion gives 102 (vs. the 109 recorded before, which included disabled/internal tools). No client-accessible tool was removed. {% endupdate %} {% update date="2026-07-30" %} ## v.1.3.0 **The biggest capability jump yet.** A major expansion of ready-to-use analysis — most of it available as one-click reports — plus higher-quality strategic insights and sturdier reliability under the hood. #### New capabilities * **16 new one-click reports** — generate a full, shareable report in a single request for: country overview, platform deep-dive, demand trends, top titles, coming soon, bundle analysis, subscription plans, awards, catalog overlap, genre distribution, price history, promotional offers, release calendar, people & talent, box office, and market intelligence. * **Hype — demand signals in the MCP** — see which titles are generating the most buzz and fold real-time demand momentum straight into your analysis and reports. * **Ask the knowledge base** — query the Origin Insights documentation directly from the assistant, without leaving your workflow. #### Quality & reliability * **More trustworthy executive insights** — the strategic bullets added to any report are now validated on the server, so out-of-spec insights are caught before they reach a report. * **Steadier under load** — longer report caching and more resilient handling of heavy query volumes keep responses fast and stable. #### Compatibility No breaking changes and nothing to update on your side. The tool surface expands to 109 tools. {% endupdate %} {% update date="2026-05-07" %} ## v.1.2.0 Capability expansion while preserving compatibility. No breaking changes in tool contracts. #### New capabilities * **Custom login** in front of OAuth — unified screen for all methods. * **Email + password login** for users who don't use Google or Microsoft, with first-login, forgot-password, and reset flows. * **Microsoft login** available as an option. * Cross-country analysis of **availability gaps** driven by demand. * **Production company exploration and analysis**: catalog, demand, reach, genres, and business model. * **Query monitoring**: traceability of what's queried and how often. #### New / changed tools | Tool | Change | | ---------------------------- | ------------------------------------------------------- | | `explore_production_company` | NEW — explore titles by studio / production company. | | `analyze_production_company` | NEW — strategic brief for a production company. | | `compare_country_catalogs` | NEW — country-by-country catalog comparison. | | `find_audience_gaps` | NEW — cross-country licensing opportunities. | | `compare_titles` | IMPROVED — adds a strategic summary for chat. | | OAuth completion | IMPROVED — direct redirect, no intermediate page. | | Auth options | IMPROVED — Google + Microsoft + email/password + reset. | | Personal API Key rotation | SAFER — requires explicit confirmation. | #### Compatibility No breaking changes. Existing OAuth clients keep working without modifications. {% endupdate %} {% update date="2026-04-30" %} ## v.1.1.0 Major productivity and reporting iteration. Cuts \~40–50% of each session's cost and fixes a query bug that broke some joins. #### New tools * `add_executive_insights` — adds 3–5 strategic points to the executive summary of any report. * `query_raw_data` — SQL escape hatch for legitimate cross-cutting queries (keeps all security gates). * `get_the_quick_pitch` — title overview in a single call (metadata + demand + availability). Replaces 3 separate calls. * `peek_behind_the_scores` — detail of the components that build each demand score. * `find_frequent_collaborators` — director / cast collaboration network. * `compare_catalog_overlap` — catalog overlap between platforms. * `discover_industry_awards` — explore award ceremonies without starting from a title. * `spot_hidden_gems` — titles with high demand and low availability. * `trace_catalog_evolution` — catalog evolution over time (improves the previous API). * `track_market_catalog_volume` — monthly volume by country and package. * `measure_title_staying_power` / `measure_season_staying_power` — days-on-platform retention. #### Visible improvements * **Auto-report**: when a response exceeds the minimum data threshold, the HTML report is generated automatically — no extra step needed. * **Auto parameter validation**: country, platform, genre, language, content type, and package type are validated with fuzzy match and return suggestions if the value is misspelled. * **Report dashboard**: `/dashboard` page with generated reports, filters, search, favorites, and auto-refresh. * **Self-service Personal API keys**: each user can check and rotate their key from `GET /me/api-key`. * **Fixed bug on JOIN queries**: some combinations returned `COLUMN_NOT_FOUND` due to an edge case in validation. Resolved. * Twice as many visualization templates (36 vs. 18). #### Efficiency | Metric | v1.0.0 | v1.1.0 | | -------------------------------------------- | ------ | --------------- | | Tool calls per data response (minimum) | 3 | 1–2 | | Tool calls per report | 5–8 | 1–2 | | Tool calls per typical session (5 questions) | \~15 | \~7–9 (-40/50%) | | {% endupdate %} | | | {% update date="2026-04-23" %} ## v.1.0.0 Production base release of the Origin Insights MCP Server: 89 tools covering search, catalog, demand, awards, prices, and reports. {% endupdate %} {% endupdates %} *** ## Coming Soon A preview of what's next for the MCP. Items graduate to the changelog once approved and shipped. * **Edit reports in place** — adjust a generated report without regenerating it: add new blocks, tweak values, and keep your changes. * **Sharper deck exports** — a quality pass on PPTX and PDF decks so every slide matches the data behind it, with cleaner charts and consistent genre naming. * **Custom reports with clearer scope** — custom reports that state exactly which filters they were built from, so there's never a doubt about what you're looking at. * **Personalized industry news** — relevant industry news filtered by your profile, turned into on-profile insights and reports. * **Talents** — ask about any cast member or director and get their titles, popularity, imagery, and current availability in one place. * **Origin Insights in ChatGPT** — a first-class ChatGPT experience for the full Origin Insights catalog. --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter: ``` GET https://knowledgebase.fabricdata.com/insights/releases/mcp.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.