diff --git a/docs/walkthrough-assets/demo-files/finance/annual-budget-forecast-2026.txt b/docs/walkthrough-assets/demo-files/finance/annual-budget-forecast-2026.txt
new file mode 100644
index 00000000..2fea8098
--- /dev/null
+++ b/docs/walkthrough-assets/demo-files/finance/annual-budget-forecast-2026.txt
@@ -0,0 +1,76 @@
+Annual Budget Forecast — Fiscal Year 2026
+
+Company: Greenleaf Technologies, Inc.
+Prepared by: Finance Department | CFO: Sarah Nakamura
+Board Presentation Date: November 12, 2025
+
+REVENUE PROJECTIONS
+
+Revenue Stream         | FY2025 Actual | FY2026 Forecast | YoY Growth
+-----------------------|---------------|-----------------|----------
+SaaS Subscriptions     |  $18,400,000  |   $24,200,000   |  +31.5%
+Enterprise Licenses    |   $6,800,000  |    $7,900,000   |  +16.2%
+Professional Services  |   $3,200,000  |    $4,100,000   |  +28.1%
+Maintenance & Support  |   $2,600,000  |    $2,900,000   |  +11.5%
+-----------------------|---------------|-----------------|----------
+Total Revenue          |  $31,000,000  |   $39,100,000   |  +26.1%
+
+COST STRUCTURE
+
+Category                   | FY2025 Actual | FY2026 Budget | % Revenue
+---------------------------|---------------|---------------|----------
+Cost of Goods Sold         |   $7,130,000  |   $8,600,000  |  22.0%
+  Cloud Infrastructure     |   $3,800,000  |   $4,500,000  |  11.5%
+  Third-Party Licenses     |   $1,200,000  |   $1,400,000  |   3.6%
+  Customer Success Team    |   $2,130,000  |   $2,700,000  |   6.9%
+                           |               |               |
+Gross Profit               |  $23,870,000  |  $30,500,000  |  78.0%
+Gross Margin               |       77.0%   |       78.0%   |
+                           |               |               |
+Operating Expenses         |               |               |
+  R&D (Engineering)        |   $9,300,000  |  $11,700,000  |  29.9%
+  Sales & Marketing        |   $7,400,000  |   $9,400,000  |  24.0%
+  General & Administrative |   $3,100,000  |   $3,500,000  |   9.0%
+Total OpEx                 |  $19,800,000  |  $24,600,000  |  62.9%
+                           |               |               |
+EBITDA                     |   $4,070,000  |   $5,900,000  |  15.1%
+EBITDA Margin              |       13.1%   |       15.1%   |
+
+HEADCOUNT PLAN
+
+Department     | Current FTEs | FY2026 Target | Net Adds
+---------------|-------------|---------------|--------
+Engineering    |     62      |      78       |   +16
+Product        |     11      |      14       |    +3
+Sales          |     24      |      32       |    +8
+Marketing      |     15      |      18       |    +3
+Customer Succ. |     18      |      23       |    +5
+G&A (HR/Fin)   |     14      |      16       |    +2
+Executive      |      6      |       6       |     0
+---------------|-------------|---------------|--------
+Total          |    150      |     187       |   +37
+
+CAPITAL EXPENDITURE
+
+Item                        | Amount      | Timeline
+----------------------------|-------------|----------
+Office expansion (Floor 3)  | $1,200,000  | Q1 2026
+Data center GPU cluster     |   $800,000  | Q2 2026
+Security infrastructure     |   $350,000  | Q1 2026
+Dev tooling & licenses      |   $280,000  | Ongoing
+Total CapEx                 | $2,630,000  |
+
+KEY ASSUMPTIONS
+1. Net revenue retention rate holds at 118% (current: 121%).
+2. Average selling price increases 8% due to product tier migration.
+3. Cloud costs grow sub-linearly due to reserved instance optimization.
+4. No acquisitions included in base forecast.
+5. USD/EUR exchange rate: 1.08 (hedged for H1 2026).
+
+RISKS & SENSITIVITIES
+  - 5% revenue shortfall → EBITDA drops to $3,950,000 (10.6% margin)
+  - 10% engineering cost overrun → EBITDA drops to $4,730,000 (12.1% margin)
+  - Delayed Floor 3 buildout → pushes 12 hires to Q3, saves $600K in H1
+
+APPROVAL SIGNATURES
+  CEO: _______________  CFO: _______________  Board Chair: _______________
diff --git a/docs/walkthrough-assets/demo-files/finance/cryptocurrency-regulatory-framework.txt b/docs/walkthrough-assets/demo-files/finance/cryptocurrency-regulatory-framework.txt
new file mode 100644
index 00000000..b0f5c88a
--- /dev/null
+++ b/docs/walkthrough-assets/demo-files/finance/cryptocurrency-regulatory-framework.txt
@@ -0,0 +1,97 @@
+Global Cryptocurrency Regulatory Framework — Comparative Analysis 2025
+
+Prepared by: Brookings Institution, Digital Assets Policy Program
+Authors: Dr. Timothy Massad, Esme Hawkins, J.D.
+Published: August 2025
+
+EXECUTIVE SUMMARY
+The global regulatory landscape for digital assets has matured significantly
+since the EU's Markets in Crypto-Assets (MiCA) regulation took full effect in
+December 2024. This paper provides a comparative analysis of cryptocurrency
+regulatory frameworks across 12 major jurisdictions, evaluating their
+approaches to consumer protection, anti-money laundering, stablecoin oversight,
+and DeFi governance.
+
+JURISDICTION COMPARISON MATRIX
+
+Jurisdiction    | Framework Status  | Stablecoin Rules | DeFi Rules | Exchange Licensing
+----------------|-------------------|------------------|------------|-------------------
+United States   | Fragmented*       | Partial (OCC)    | Pending    | State-by-state
+European Union  | Comprehensive     | Full (MiCA)      | Under study| Mandatory (CASP)
+United Kingdom  | Phased rollout    | Proposed          | Sandbox    | FCA registration
+Japan           | Comprehensive     | Full (PSA)       | Guidelines | Mandatory (FSA)
+Singapore       | Comprehensive     | Full (PS Act)    | Risk-based | Mandatory (MAS)
+Switzerland     | Comprehensive     | Integrated       | Principles | Mandatory (FINMA)
+UAE             | Comprehensive     | Full (VARA)      | Permitted  | Mandatory (VARA)
+Hong Kong       | Phased rollout    | Proposed          | Pending    | Mandatory (SFC)
+South Korea     | Implementation    | Proposed          | Pending    | Mandatory
+India           | Tax framework     | Banned            | Banned     | Under consideration
+China           | Prohibition       | CBDC only         | Banned     | Banned
+Brazil          | Framework law     | Under development | Pending    | Mandatory (BCB)
+
+* US: SEC, CFTC, OCC, FinCEN, state regulators all have overlapping jurisdiction.
+  FIT21 (Financial Innovation and Technology for the 21st Century Act) passed
+  House in 2024; Senate debate ongoing as of Aug 2025.
+
+STABLECOIN REGULATION DEEP DIVE
+
+Global stablecoin market cap: $178 billion (Aug 2025)
+  - USDT (Tether): $95B (53%)
+  - USDC (Circle): $52B (29%)
+  - DAI (MakerDAO): $8.2B (5%)
+  - Others: $22.8B (13%)
+
+MiCA Requirements (EU):
+  - 1:1 reserve backing in high-quality liquid assets (cash, government bonds)
+  - Reserves held at EU credit institutions (30% minimum in cash)
+  - Quarterly reserve attestations by independent auditors
+  - €200M daily transaction cap for non-euro stablecoins (significant asset-
+    referenced tokens)
+  - Tether (USDT): Non-compliant as of June 2025; delisted from EU exchanges
+
+US Approach:
+  - OCC interpretive letter (Jan 2025): National banks may issue stablecoins
+  - Lummis-Gillibrand Payment Stablecoin Act: Requires 100% reserves, monthly
+    attestations, $10B issuer threshold for Fed oversight
+  - Status: Committee markup completed; floor vote expected Q4 2025
+
+DECENTRALIZED FINANCE (DeFi) CHALLENGES
+
+Total Value Locked (DeFi): $148 billion (Aug 2025)
+Top protocols: Lido ($38B), Aave ($22B), MakerDAO ($18B), Uniswap ($11B)
+
+Regulatory questions unresolved:
+1. Who is the "issuer" or "operator" of a decentralized protocol?
+2. Can smart contract code constitute a "prospectus" or "white paper"?
+3. How do KYC/AML requirements apply to non-custodial protocols?
+4. Jurisdiction over DAOs with globally distributed token holders?
+
+EU approach: European Commission study (due Dec 2025) evaluating whether MiCA
+requires amendment for DeFi. Likely outcome: "same activity, same risk, same
+regulation" with implementation guidance for protocol operators.
+
+ENFORCEMENT STATISTICS (2023–2025)
+
+Agency    | Actions Filed | Penalties Assessed | Key Cases
+----------|---------------|--------------------|-----------
+SEC (US)  |      78       |     $4.2B          | Coinbase, Binance
+CFTC (US) |      34       |     $1.8B          | FTX settlement
+DOJ (US)  |      22       |     $3.1B          | Binance (CZ plea)
+FCA (UK)  |      15       |     £180M          | Unregistered exchanges
+MAS (SG)  |       8       |     S$45M          | 3AC-related entities
+
+CONSUMER PROTECTION METRICS
+
+Estimated crypto fraud losses (global, 2024): $14.8 billion
+  - Pig-butchering scams: $4.2B (28%)
+  - Rug pulls / exit scams: $3.1B (21%)
+  - Phishing / social engineering: $2.8B (19%)
+  - Ponzi / yield schemes: $2.4B (16%)
+  - Exchange hacks: $2.3B (16%)
+
+POLICY RECOMMENDATIONS
+1. Establish international stablecoin standards through FSB/IOSCO coordination.
+2. Adopt activity-based (not entity-based) regulation for DeFi protocols.
+3. Create regulatory sandboxes for tokenized real-world assets (RWA).
+4. Mandate on-chain analytics for AML compliance (Chainalysis, Elliptic).
+5. Harmonize tax reporting (OECD Crypto-Asset Reporting Framework by 2027).
diff --git a/docs/walkthrough-assets/demo-files/finance/q3-2025-portfolio-performance-report.txt b/docs/walkthrough-assets/demo-files/finance/q3-2025-portfolio-performance-report.txt
new file mode 100644
index 00000000..f1cc92e0
--- /dev/null
+++ b/docs/walkthrough-assets/demo-files/finance/q3-2025-portfolio-performance-report.txt
@@ -0,0 +1,72 @@
+Quarterly Portfolio Performance Report — Q3 2025
+
+Prepared for: Meridian Capital Partners, LP
+Investment Manager: Crestline Wealth Advisors
+Date: October 15, 2025
+
+PORTFOLIO SUMMARY
+
+Total AUM (as of Sep 30, 2025):  $48,720,000
+Net Return (Q3):                 +4.82%
+Benchmark (S&P 500 TR):         +3.61%
+Alpha Generated:                 +1.21%
+Sharpe Ratio (trailing 12-mo):  1.34
+
+ASSET ALLOCATION
+
+Asset Class             | Allocation | Q3 Return | Benchmark
+------------------------|------------|-----------|----------
+US Large-Cap Equity     |    35.0%   |   +5.12%  |  +4.20%
+US Mid/Small-Cap Equity |    12.0%   |   +6.84%  |  +5.30%
+International Developed |    15.0%   |   +3.45%  |  +2.90%
+Emerging Markets        |     8.0%   |   +2.18%  |  +1.60%
+Investment-Grade Bonds  |    15.0%   |   +1.92%  |  +1.75%
+High-Yield Bonds        |     5.0%   |   +3.28%  |  +2.90%
+Real Assets (REITs)     |     5.0%   |   +7.62%  |  +6.40%
+Alternatives (PE/HF)    |     3.0%   |   +4.10%  |    N/A
+Cash & Equivalents      |     2.0%   |   +1.32%  |  +1.30%
+
+TOP EQUITY HOLDINGS
+
+Symbol  | Name                  | Weight | Q3 Return | Contribution
+--------|-----------------------|--------|-----------|-------------
+NVDA    | NVIDIA Corporation    |  4.2%  |  +18.3%   |   +0.77%
+MSFT    | Microsoft Corp        |  3.8%  |   +7.1%   |   +0.27%
+AAPL    | Apple Inc             |  3.5%  |   +4.6%   |   +0.16%
+AMZN    | Amazon.com Inc        |  3.1%  |   +9.2%   |   +0.29%
+LLY     | Eli Lilly & Co        |  2.7%  |  +12.4%   |   +0.33%
+UNH     | UnitedHealth Group    |  2.3%  |   +5.8%   |   +0.13%
+JPM     | JPMorgan Chase        |  2.1%  |   +8.9%   |   +0.19%
+V       | Visa Inc              |  1.9%  |   +6.3%   |   +0.12%
+AVGO    | Broadcom Inc          |  1.8%  |  +14.7%   |   +0.26%
+XOM     | Exxon Mobil Corp      |  1.6%  |   -2.1%   |   -0.03%
+
+RISK METRICS
+
+  Portfolio Beta:                0.92
+  Max Drawdown (Q3):           -3.14% (Aug 5 market selloff)
+  VaR (95%, 1-day):            $243,600
+  Tracking Error (annualized):  2.18%
+  Information Ratio:            0.55
+
+COMMENTARY
+The portfolio outperformed the benchmark by 121 basis points in Q3, driven
+primarily by our overweight in semiconductor and healthcare names. The August
+volatility event (triggered by Bank of Japan rate concerns) was managed through
+our systematic rebalancing discipline—we added to equities on the drawdown,
+which contributed approximately 40 bps of alpha recovery.
+
+Our alternatives sleeve delivered steady returns from the Sequoia Growth Fund IV
+(vintage 2022), which marked up several portfolio companies after strong Q2
+earnings. The real assets allocation benefited from falling interest rate
+expectations, with REITs posting the strongest quarterly return across our sleeve.
+
+OUTLOOK & RECOMMENDED ACTIONS
+1. Reduce US Large-Cap Equity by 2% (take profits on semiconductor gains).
+2. Increase International Developed by 2% (European value opportunity).
+3. Maintain fixed-income duration at 5.2 years (anticipating Fed rate cuts).
+4. Review high-yield exposure for credit quality deterioration signals.
+
+DISCLOSURES
+Past performance does not guarantee future results. This report is confidential
+and intended solely for the named recipient.
diff --git a/docs/walkthrough-assets/demo-files/finance/venture-capital-term-sheet.txt b/docs/walkthrough-assets/demo-files/finance/venture-capital-term-sheet.txt
new file mode 100644
index 00000000..565f1797
--- /dev/null
+++ b/docs/walkthrough-assets/demo-files/finance/venture-capital-term-sheet.txt
@@ -0,0 +1,91 @@
+SERIES B PREFERRED STOCK TERM SHEET
+
+Company:          NovaBridge AI, Inc. (Delaware C-Corp)
+Investors:        Horizon Ventures (Lead), Catalyst Capital, Redwood Partners
+Round Size:       $35,000,000
+Pre-Money Valuation: $140,000,000
+Post-Money Valuation: $175,000,000
+Date:             August 20, 2025
+Status:           Non-Binding (except Confidentiality, Exclusivity, Expenses)
+
+SECURITIES AND PRICING
+
+  Security: Series B Preferred Stock
+  Price per share: $8.75 (based on 16,000,000 Series B shares)
+  Authorized shares post-round: 20,000,000 Series B Preferred
+
+INVESTMENT AMOUNTS
+
+  Horizon Ventures:    $20,000,000  (2,285,714 shares)
+  Catalyst Capital:    $10,000,000  (1,142,857 shares)
+  Redwood Partners:     $5,000,000  (  571,429 shares)
+
+CAPITALIZATION (Post-Closing)
+
+  Founders & Common:     10,000,000 shares  (50.0%)
+  Series A Preferred:     4,000,000 shares  (20.0%)
+  Series B Preferred:     4,000,000 shares  (20.0%)
+  ESOP (unallocated):     2,000,000 shares  (10.0%)
+  Total Fully Diluted:   20,000,000 shares
+
+LIQUIDATION PREFERENCE
+  1× non-participating preferred. Series B has senior liquidation preference
+  over Series A and Common. Upon a Deemed Liquidation Event, Series B holders
+  receive the greater of (i) 1× their Original Purchase Price plus declared
+  but unpaid dividends, or (ii) their pro-rata share on an as-converted basis.
+
+ANTI-DILUTION
+  Broad-based weighted average anti-dilution protection.
+
+DIVIDENDS
+  Non-cumulative, 6% per annum when and if declared by the Board.
+  Participation pari passu with Common on an as-converted basis.
+
+CONVERSION
+  Automatic conversion to Common upon (a) IPO with gross proceeds ≥ $75M and
+  per-share price ≥ 3× Series B price, or (b) written consent of 60% of
+  Series B holders. Optional conversion at any time at holder's election.
+
+VOTING RIGHTS
+  One vote per share on an as-converted basis. Series B votes as a separate
+  class on: (i) amendments to Series B terms, (ii) issuance of senior or
+  pari passu securities, (iii) dividends or redemptions, (iv) changes to
+  Board size.
+
+BOARD COMPOSITION
+  5 directors:
+    2 — Common holders (Founders)
+    2 — Series B (1 Horizon, 1 Catalyst)
+    1 — Independent (mutually agreed)
+
+PROTECTIVE PROVISIONS (Series B Consent Required)
+  - Any liquidation, merger, or sale of substantially all assets
+  - Debt exceeding $2,000,000
+  - Related-party transactions > $250,000
+  - Changes to authorized share count
+  - Annual budget approval (if not adopted by Dec 31)
+
+RIGHT OF FIRST REFUSAL & CO-SALE
+  Pro-rata right of first refusal on new issuances. Co-sale right on
+  founder transfers exceeding 5% of holdings per 12-month period.
+
+DRAG-ALONG
+  If holders of 60% of Preferred and a majority of Common approve a sale,
+  all stockholders must participate on the same terms.
+
+INFORMATION RIGHTS
+  Monthly financial statements (within 30 days), annual audited financials
+  (within 90 days), annual budget. Inspection rights with 5 business days notice.
+
+VESTING
+  Founder shares: 4-year vesting, 1-year cliff, monthly thereafter.
+  Acceleration: 50% single-trigger on Change of Control; 100% double-trigger.
+
+EXCLUSIVITY
+  45 days from execution of this term sheet.
+
+EXPENSES
+  Company to reimburse lead investor legal fees up to $75,000.
+
+GOVERNING LAW
+  State of Delaware.
diff --git a/docs/walkthrough-assets/demo-files/mixed/ai-semiconductor-industry-analysis.txt b/docs/walkthrough-assets/demo-files/mixed/ai-semiconductor-industry-analysis.txt
new file mode 100644
index 00000000..95367e77
--- /dev/null
+++ b/docs/walkthrough-assets/demo-files/mixed/ai-semiconductor-industry-analysis.txt
@@ -0,0 +1,85 @@
+AI Semiconductor Industry Analysis — Market Outlook 2025–2030
+
+Morgan Stanley Equity Research | Sector: Technology Hardware
+Lead Analyst: Jennifer Park, CFA
+Date: September 2025
+
+INVESTMENT THESIS
+The AI accelerator market is entering a sustained growth phase driven by
+enterprise AI adoption, sovereign AI infrastructure investment, and the
+scaling requirements of frontier model training. We project the total
+addressable market to grow from $68B (2024) to $320B by 2030, a 29% CAGR.
+
+MARKET SEGMENTATION (2024 → 2030E)
+
+Segment              | 2024 Revenue | 2030E Revenue | CAGR
+---------------------|-------------|---------------|------
+Data Center GPU      |   $47.5B    |   $185B       |  25%
+Custom ASIC (TPU etc)|    $8.2B    |    $62B       |  40%
+Edge AI Chips        |    $6.8B    |    $38B       |  33%
+HBM Memory           |    $5.5B    |    $35B       |  36%
+
+COMPETITIVE LANDSCAPE
+
+NVIDIA (NVDA) — Dominant Incumbent
+  Market share (DC GPU): 88% (2024)
+  Key product: Blackwell B200 ($30–40K ASP)
+  Moat: CUDA ecosystem (4M+ developers), networking (Spectrum-X)
+  Risk: Customer ASIC development (Google TPU, Amazon Trainium)
+  Rating: Overweight | PT: $185
+
+AMD (AMD) — Credible Challenger
+  Market share (DC GPU): 9% (2024)
+  Key product: MI350X (launching Q1 2026)
+  Opportunity: ROCm 7 closing software gap; hyperscaler design wins
+  Rating: Equal Weight | PT: $175
+
+Intel (INTC) — Restructuring
+  Market share (DC GPU): <2%
+  Key product: Gaudi 3 (competitive TCO but limited adoption)
+  Challenge: Foundry losses, management transition
+  Rating: Underweight | PT: $22
+
+Broadcom (AVGO) — Custom ASIC Leader
+  Revenue: $4.2B from AI ASIC (Google, Meta, ByteDance)
+  Opportunity: 3 additional hyperscaler engagements expected by 2026
+  Rating: Overweight | PT: $230
+
+SUPPLY CHAIN ANALYSIS
+
+TSMC remains the critical bottleneck:
+  - CoWoS advanced packaging capacity: ~35K wafers/month (2025)
+  - Expanding to ~80K wafers/month by end of 2026
+  - N3E (3nm) allocation for AI: 60% of total wafer starts
+
+HBM (High Bandwidth Memory) suppliers:
+  SK Hynix: 53% market share (HBM3E), sole supplier for NVIDIA B200
+  Samsung: 38% market share, qualified for AMD MI350X
+  Micron: 9% share, ramping HBM3E yield improvements
+
+SOVEREIGN AI INVESTMENT WAVE
+
+Country/Region    | Announced Investment | Timeline | Key Projects
+------------------|---------------------|----------|---------------
+Saudi Arabia      |     $100B           | 2025–30  | NEOM AI campus
+UAE               |      $30B           | 2025–28  | G42 data centers
+France            |      $25B           | 2025–30  | "AI France" plan
+Japan             |      $13B           | 2025–27  | NVIDIA Blackwell cluster
+India             |      $10B           | 2025–28  | IndiaAI Mission
+
+RISKS TO THESIS
+1. Export controls: Further US restrictions on China could reduce NVIDIA TAM
+   by ~$15B annually (already reflected in our base case).
+2. Model scaling plateau: If scaling laws diminish, training CAPEX could
+   decelerate faster than inference growth compensates.
+3. Power constraints: 1 GW+ data centers require utility-scale power; permitting
+   delays could push installations out 12–18 months.
+
+VALUATION SUMMARY
+  NVDA: 35× FY2027E EPS = $185 PT (20% upside)
+  AMD:  30× FY2027E EPS = $175 PT (8% upside)
+  AVGO: 28× FY2027E EPS = $230 PT (15% upside)
+  INTC: 18× FY2027E EPS = $22 PT (12% downside)
+
+DISCLAIMER: This analysis is for informational purposes only and does not
+constitute investment advice. Past performance is not indicative of future results.
diff --git a/docs/walkthrough-assets/demo-files/mixed/amazon-deforestation-economic-impact.txt b/docs/walkthrough-assets/demo-files/mixed/amazon-deforestation-economic-impact.txt
new file mode 100644
index 00000000..677551ee
--- /dev/null
+++ b/docs/walkthrough-assets/demo-files/mixed/amazon-deforestation-economic-impact.txt
@@ -0,0 +1,43 @@
+Amazon Deforestation and Its Economic Impact on Global Carbon Markets
+
+World Resources Institute | Policy Brief No. 42, July 2025
+
+SUMMARY
+Between 2001 and 2024, the Brazilian Amazon lost approximately 230,000 km² of
+primary forest—an area larger than the United Kingdom. This policy brief
+quantifies the economic externalities of continued deforestation through the
+lens of carbon credit markets, agricultural commodity supply chains, and
+ecosystem service valuation.
+
+CARBON ECONOMICS
+  Carbon stored in Amazon biomass: ~150 billion tonnes CO₂e
+  Annual emissions from deforestation (2020–2024 avg): 1.1 Gt CO₂e
+  Social Cost of Carbon (US EPA, 2025): $230/tonne CO₂e
+  Implied annual externality: $253 billion
+
+The voluntary carbon market valued Amazonian REDD+ credits at $8–15/tonne in
+2024, a fraction of the social cost. The EU Carbon Border Adjustment Mechanism
+(CBAM), effective 2026, will increase the effective price of embedded
+deforestation carbon in agricultural exports to the EU to ~€90/tonne.
+
+AGRICULTURAL TRADE IMPACTS
+Brazil exported $47B in soybeans and $12B in beef in 2024. The EU Deforestation
+Regulation (EUDR) requires supply-chain traceability to plot-level coordinates
+by January 2026. Compliance costs are estimated at $1.2B for the Brazilian
+agricultural sector, but non-compliance risks $6.8B in annual EU market access.
+
+ECOSYSTEM SERVICE VALUATION
+  Service                  | Value ($/ha/year) | Total Amazon Value
+  -------------------------|-------------------|-------------------
+  Carbon sequestration     |     $1,800        |  $630 billion
+  Water cycling            |       $720        |  $252 billion
+  Biodiversity (existence) |       $340        |  $119 billion
+  Pollination services     |       $180        |   $63 billion
+  Flood regulation         |       $150        |   $53 billion
+
+RECOMMENDATIONS
+1. Align carbon credit pricing with social cost of carbon through compliance
+   market integration.
+2. Support Brazil's PPCDAm enforcement with $500M multilateral funding.
+3. Accelerate EUDR implementation with technical assistance for smallholders.
+4. Invest in high-resolution satellite monitoring (Planet, NICFI partnership).
diff --git a/docs/walkthrough-assets/demo-files/nature/arctic-sea-ice-monitoring-report.txt b/docs/walkthrough-assets/demo-files/nature/arctic-sea-ice-monitoring-report.txt
new file mode 100644
index 00000000..7814b653
--- /dev/null
+++ b/docs/walkthrough-assets/demo-files/nature/arctic-sea-ice-monitoring-report.txt
@@ -0,0 +1,70 @@
+Arctic Sea Ice Monitoring Report — September 2025 Minimum
+
+National Snow and Ice Data Center (NSIDC)
+Authors: Dr. Mark Serreze, Dr. Julienne Stroeve
+Release Date: October 1, 2025
+
+ANNUAL MINIMUM EXTENT
+The 2025 Arctic sea ice minimum extent was recorded on September 14 at
+4.08 million km², ranking as the 4th lowest in the 46-year satellite record
+(1979–2025).
+
+RANKING OF LOWEST SEPTEMBER MINIMUMS
+  1. 2012:  3.39 million km²
+  2. 2020:  3.74 million km²
+  3. 2019:  4.01 million km²
+  4. 2025:  4.08 million km²  ← This year
+  5. 2023:  4.12 million km²
+
+The 1981–2010 median September minimum was 6.30 million km². The 2025 value
+represents a 35% departure from this climatological baseline.
+
+ICE THICKNESS AND VOLUME
+CryoSat-2 and ICESat-2 altimetry indicate mean ice thickness of 1.24 m across
+the Arctic basin (September average), continuing the long-term thinning trend
+of -0.13 m/decade. Estimated total ice volume: 4,150 km³ (PIOMAS model),
+compared to ~12,000 km³ in the early 1980s.
+
+REGIONAL BREAKDOWN
+
+Region              | Sep 2025 Extent | Anomaly vs. Median | Trend
+--------------------|-----------------|--------------------|---------
+Beaufort Sea        |   310,000 km²   |    -42%            | -14%/dec
+Chukchi Sea         |   195,000 km²   |    -51%            | -18%/dec
+East Siberian Sea   |   280,000 km²   |    -38%            | -12%/dec
+Laptev Sea          |   125,000 km²   |    -55%            | -20%/dec
+Kara Sea            |   165,000 km²   |    -33%            | -11%/dec
+Barents Sea         |    48,000 km²   |    -61%            | -22%/dec
+Central Arctic      | 2,960,000 km²   |    -22%            |  -8%/dec
+
+CONTRIBUTING FACTORS
+
+1. Atmospheric Circulation: A persistent Arctic dipole pattern (high pressure
+   over Beaufort, low over Eurasia) promoted ice transport out of the Transpolar
+   Drift Stream through Fram Strait.
+
+2. Ocean Heat: Atlantic Water inflow through the Barents Sea Opening was 15%
+   above the 2000–2020 mean, contributing to early melt onset in the Barents
+   and Kara Seas (May 8 vs. median May 22).
+
+3. Albedo Feedback: Early melt pond formation (detected May 15 via MODIS)
+   reduced surface albedo from 0.82 to 0.45, amplifying solar absorption
+   during the 24-hour daylight period.
+
+IMPLICATIONS FOR WILDLIFE
+  - Polar bear (Ursus maritimus): Extended ice-free season in western Hudson
+    Bay reached 168 days (record; mean 135 days), increasing nutritional stress.
+  - Walrus (Odobenus rosmarus): Haul-out events on Alaskan coast began August 2
+    (2 weeks earlier than 2010–2020 mean).
+  - Arctic cod (Boreogadus saida): Under-ice habitat area reduced ~30%;
+    northward range shift of 200 km observed in acoustic surveys.
+
+SHIPPING AND GEOPOLITICAL NOTES
+The Northern Sea Route was ice-free for transit without icebreaker support from
+July 28 to October 12 (77 days), the second-longest open window recorded. Total
+transits in 2025: 127 vessels (up from 92 in 2024).
+
+FORECAST
+NSIDC seasonal forecast models project a 30–40% probability that September
+minimum extent will fall below 3.5 million km² within the next 5 years, and a
+>50% probability of an effectively ice-free Arctic (<1 million km²) by 2040.
diff --git a/docs/walkthrough-assets/demo-files/nature/coral-reef-health-assessment.txt b/docs/walkthrough-assets/demo-files/nature/coral-reef-health-assessment.txt
new file mode 100644
index 00000000..e22c7217
--- /dev/null
+++ b/docs/walkthrough-assets/demo-files/nature/coral-reef-health-assessment.txt
@@ -0,0 +1,55 @@
+Coral Reef Health Assessment — Indo-Pacific Survey 2025
+
+Marine Biological Laboratory | Woods Hole Oceanographic Institution
+Lead: Dr. Anika Chowdhury, Reef Systems Group
+Survey Period: March 15 – June 30, 2025
+
+EXECUTIVE SUMMARY
+We conducted 312 belt-transect surveys across 14 reef sites in the Coral Triangle
+region (Indonesia, Philippines, Papua New Guinea). Overall live coral cover averaged
+38.2% (±6.7), a modest recovery from the 2023 bleaching event low of 31.4%.
+
+SITE HIGHLIGHTS
+
+Raja Ampat, Indonesia
+  Live coral cover: 52.1%  |  Bleaching prevalence: 4.2%
+  Dominant genera: Acropora, Porites, Montipora
+  Fish biomass: 184 kg/ha (above regional mean)
+  Notes: Marine Protected Area enforcement strong; illegal blast fishing
+  incidents down 60% year-over-year
+
+Tubbataha Reefs, Philippines
+  Live coral cover: 41.8%  |  Bleaching prevalence: 8.1%
+  Dominant genera: Pocillopora, Porites
+  Fish biomass: 127 kg/ha
+  Notes: Crown-of-thorns starfish (COTS) outbreak contained via diver removal
+  program; 2,300 COTS removed Q2 2025
+
+Kimbe Bay, Papua New Guinea
+  Live coral cover: 29.4%  |  Bleaching prevalence: 12.6%
+  Dominant genera: Porites, Pavona
+  Fish biomass: 89 kg/ha (below threshold)
+  Notes: Sedimentation from upstream logging remains primary stressor
+
+METHODOLOGY
+  Transect protocol: 50 m × 2 m belt, point-intercept at 50 cm intervals
+  Depth range: 3–15 m
+  Photo quadrats: 0.5 m² per station, 4 stations per transect
+  Water quality: YSI EXO2 sonde (temperature, salinity, turbidity, DO, pH)
+
+THERMAL STRESS INDEX
+  Degree Heating Weeks (DHW) peaked at 6.2 in southern Philippines (April 2025),
+  triggering Alert Level 1. Raja Ampat remained below 3.0 DHW throughout the survey.
+
+RECOMMENDATIONS
+1. Expand COTS early-detection buoys to 6 additional Philippines sites.
+2. Establish sediment traps in Kimbe Bay watershed; coordinate with forestry ministry.
+3. Deploy autonomous reef-monitoring cameras (ARMS units) at all 14 sites for
+   year-round benthic cover tracking.
+4. Prioritize genetic sampling of heat-tolerant Acropora colonies for assisted gene
+   flow pilot program.
+
+APPENDIX
+  Full transect data: reef_survey_2025_transects.xlsx
+  Photo archive: /reef_photos/2025_IndoPacific/
+  Water quality time series: wq_timeseries_mar_jun2025.csv
diff --git a/docs/walkthrough-assets/demo-files/nature/monarch-butterfly-migration-study.txt b/docs/walkthrough-assets/demo-files/nature/monarch-butterfly-migration-study.txt
new file mode 100644
index 00000000..695307fb
--- /dev/null
+++ b/docs/walkthrough-assets/demo-files/nature/monarch-butterfly-migration-study.txt
@@ -0,0 +1,49 @@
+Monarch Butterfly Migration Patterns — Field Report 2025
+
+Principal Investigator: Dr. Elena Vasquez, Department of Entomology
+Location: Pacific Grove Monarch Sanctuary, California
+Season: October 2025 — February 2026
+
+ABSTRACT
+This report documents the annual overwintering aggregation of Danaus plexippus
+(monarch butterfly) at the Pacific Grove sanctuary. Our team recorded an estimated
+24,600 individuals—a 14% increase over the 2024 season—clustered primarily in
+Monterey pine and eucalyptus groves within a 2.3-hectare core area.
+
+KEY OBSERVATIONS
+
+Temperature & Microclimate
+  Average canopy temperature: 12.4 °C (±1.1)
+  Average humidity: 78% RH
+  Wind speed at cluster height: < 2 m/s
+
+Population Dynamics
+  Peak arrival: November 8–14
+  Peak departure: February 18–22
+  Mortality events: 2 frost events (Jan 3, Jan 17) caused ~3% attrition
+
+Tagging Results
+  Tags deployed: 1,200 (wing-adhesive alphanumeric)
+  Re-sights from northern sites: 47 individuals traced to Oregon/Washington
+  Longest confirmed flight: 2,840 km (Ashland, OR → Pacific Grove, CA)
+
+HABITAT NOTES
+The primary roosting trees—Monterey pines aged 60–80 years—showed signs of pitch
+canker (Fusarium circinatum). Three mature trees were removed after the 2024 season,
+reducing canopy cover by ~8%. Supplemental eucalyptus plantings (2022 cohort) have
+reached 6 m height and are beginning to provide viable alternative roost sites.
+
+Milkweed (Asclepias fascicularis) restoration plots along the coastal corridor showed
+a 22% increase in stem density, which correlates with higher early-season oviposition
+rates observed in the central valley waypoints.
+
+CONSERVATION RECOMMENDATIONS
+1. Continue annual Fusarium monitoring and selective pruning.
+2. Expand native milkweed plantings at highway median strips (Hwy 1 corridor).
+3. Install additional weather stations at 5 m and 15 m canopy heights.
+4. Partner with citizen-science platform iNaturalist for broader re-sight data.
+
+DATA FILES
+  Raw tag data: monarch_tags_2025.csv
+  Weather logs: pacific_grove_wx_oct25_feb26.json
+  Photo transects: /transects/PG_2025_T01–T48/
diff --git a/docs/walkthrough-assets/demo-files/nature/yellowstone-wolf-reintroduction-update.txt b/docs/walkthrough-assets/demo-files/nature/yellowstone-wolf-reintroduction-update.txt
new file mode 100644
index 00000000..b058a731
--- /dev/null
+++ b/docs/walkthrough-assets/demo-files/nature/yellowstone-wolf-reintroduction-update.txt
@@ -0,0 +1,64 @@
+Yellowstone Wolf Reintroduction — 30-Year Ecological Update
+
+U.S. Fish and Wildlife Service  |  National Park Service
+Authors: Dr. James Whitfield, Dr. Rosa Delgado
+Report Date: September 2025
+
+OVERVIEW
+Thirty years after the landmark 1995 reintroduction of gray wolves (Canis lupus) to
+Yellowstone National Park, the ecosystem continues to demonstrate cascading ecological
+benefits. The 2025 winter census confirmed 108 wolves in 10 packs across the Greater
+Yellowstone Ecosystem (GYE).
+
+PACK CENSUS (Winter 2025)
+
+Pack Name          | Territory          | Adults | Pups | Total
+-------------------|--------------------|--------|------|------
+Lamar Canyon       | Lamar Valley       |    9   |   4  |  13
+Junction Butte     | Slough Creek       |   11   |   5  |  16
+Mollie's           | Pelican Valley     |    8   |   3  |  11
+Wapiti Lake        | Hayden Valley      |    7   |   2  |   9
+8-Mile             | Northern Range     |   10   |   4  |  14
+Prospect Peak      | Mirror Plateau     |    5   |   2  |   7
+Canyon              | Canyon Village     |    6   |   3  |   9
+Bechler            | SW Yellowstone     |    7   |   3  |  10
+Snake River        | Southern Range     |    6   |   2  |   8
+Phantom Lake       | Central Plateau    |    8   |   3  |  11
+
+TROPHIC CASCADE METRICS
+
+Elk (Cervus canadensis) Population
+  Northern herd estimate: 6,200 (down from ~20,000 pre-wolf)
+  Calf-to-cow ratio: 28:100 (healthy; pre-wolf was 45:100 with overgrazing)
+  Behavioral shift: Elk spend 60% less time in riparian zones vs. pre-wolf era
+
+Vegetation Recovery
+  Willow (Salix spp.) canopy height: +340% in Lamar Valley since 1995
+  Aspen (Populus tremuloides) recruitment: measurable in 68% of monitored stands
+  Cottonwood regeneration: observed in 12 of 15 permanent plots along rivers
+
+Beaver (Castor canadensis) Response
+  Active beaver colonies in northern Yellowstone: 12 (zero in 1995)
+  Dam count (Slough Creek drainage): 23
+  Pond surface area created: ~4.8 hectares
+
+Scavenger Guild
+  Raven, magpie, bald eagle, and grizzly bear all benefit from wolf-killed
+  carcasses during winter. Carcass availability index: 3.2× pre-wolf levels.
+
+HUMAN DIMENSIONS
+  Wolf-livestock depredation claims (GYE, 2024): 18 confirmed cattle, 6 sheep
+  Compensation paid: $127,400 (Defenders of Wildlife fund)
+  Wolf-watching tourism economic impact: est. $82 million/year for gateway communities
+
+GENETIC HEALTH
+  Heterozygosity (He): 0.71 (stable; founder population He = 0.74)
+  Effective population size (Ne): ~64
+  Recommendation: Evaluate genetic connectivity with Idaho/Montana meta-population
+  via corridor analysis.
+
+FUTURE PRIORITIES
+1. Expand GPS collar deployment to 30% of adults for fine-scale territory mapping.
+2. Continue long-term vegetation plots (30+ year dataset now globally significant).
+3. Model climate-change impacts on elk migration timing and wolf prey availability.
+4. Coordinate with state agencies on buffer-zone management outside park boundaries.
diff --git a/docs/walkthrough-assets/demo-files/research/crispr-gene-therapy-clinical-trials.txt b/docs/walkthrough-assets/demo-files/research/crispr-gene-therapy-clinical-trials.txt
new file mode 100644
index 00000000..6aa3cf5d
--- /dev/null
+++ b/docs/walkthrough-assets/demo-files/research/crispr-gene-therapy-clinical-trials.txt
@@ -0,0 +1,115 @@
+CRISPR-Based Gene Therapy: Clinical Trial Landscape and Regulatory Pathways
+
+Authors: Dr. Maya Okonkwo, Dr. Liam Foster, Dr. Kenji Tanaka
+Affiliation: Center for Genomic Medicine, Massachusetts General Hospital
+Published: Annual Review of Genomics and Human Genetics, Vol. 26, 2025
+
+ABSTRACT
+CRISPR-Cas9 and its next-generation variants (base editors, prime editors) have
+advanced from laboratory tools to clinical therapeutics at unprecedented speed.
+As of mid-2025, 87 clinical trials involving CRISPR-based interventions are
+registered globally, spanning oncology, hematology, ophthalmology, and rare
+genetic disorders. This review catalogs the current trial landscape, analyzes
+safety and efficacy data from completed Phase I/II studies, and examines the
+evolving regulatory frameworks in the US, EU, and China.
+
+1. APPROVED THERAPIES
+
+Casgevy (exagamglogene autotemcel) — Vertex/CRISPR Therapeutics
+  Indication: Sickle cell disease (SCD) and transfusion-dependent β-thalassemia
+  Mechanism: Ex vivo editing of BCL11A enhancer in autologous HSCs
+  Approval: FDA (Dec 2023), EMA (Feb 2024), MHRA (Nov 2023)
+  Outcomes (36-month follow-up):
+    SCD: 97% of patients (29/30) free of vaso-occlusive crises
+    β-thal: 93% of patients (28/30) transfusion-independent
+  Safety: Myeloablative conditioning (busulfan) remains primary risk factor
+
+2. PHASE III TRIALS (Selected)
+
+Trial ID     | Sponsor       | Indication         | Editor    | Status
+-------------|---------------|--------------------|-----------|---------
+NCT06128252  | Editas Med    | LCA10 (blindness)  | Cas9      | Enrolling
+NCT05951205  | Intellia      | ATTR amyloidosis   | Cas9      | Phase III
+NCT06234100  | Beam Ther.    | SCD (base edit)    | ABE8e     | Phase II/III
+NCT06301920  | Prime Med.    | Wilson disease      | PE2       | Phase II
+NCT06445500  | Caribou Bio   | r/r B-cell NHL     | Cas12a    | Phase II
+
+3. IN VIVO vs. EX VIVO APPROACHES
+
+                    | Ex Vivo                  | In Vivo
+--------------------|--------------------------|---------------------------
+Delivery            | Electroporation of cells | LNP or AAV to target organ
+Editing efficiency  | >90% (controlled)        | 30–70% (tissue-dependent)
+Off-target control  | High (screened pre-infusion)| Lower (limited monitoring)
+Cost per patient    | $1.5–2.5M                | Est. $500K–1M
+Scalability         | Limited (autologous)     | Higher (off-the-shelf)
+
+Intellia's NTLA-2001 (in vivo liver CRISPR for transthyretin amyloidosis)
+demonstrated 93% reduction in serum TTR protein at 2 years—the strongest
+in vivo editing result to date. This has catalyzed a wave of liver-targeted
+programs using lipid nanoparticle (LNP) delivery.
+
+4. SAFETY LANDSCAPE
+
+4.1 Off-Target Editing
+  GUIDE-seq and CIRCLE-seq nominated <5 off-target sites per therapeutic
+  gRNA. No clinical adverse events attributable to off-target editing have
+  been reported through 2025.
+
+4.2 Genotoxicity
+  Large deletions at on-target sites (>1 kb) detected at 1–5% frequency
+  in preclinical studies. Clinical significance unclear; long-term follow-up
+  (15 years per FDA mandate) ongoing.
+
+4.3 Immune Response
+  Anti-Cas9 antibodies detected in 58% of patients (pre-existing immunity
+  from S. pyogenes exposure). Clinical impact: negligible for ex vivo
+  approaches; potential concern for repeat in vivo dosing.
+
+4.4 Reported Serious Adverse Events (all CRISPR trials, 2019–2025)
+  - Myeloablative conditioning toxicity: 12 events (expected)
+  - Cytokine release syndrome (CAR-T + CRISPR): 8 events (Grade 3+)
+  - Editing-related SAEs: 0 confirmed
+
+5. NEXT-GENERATION EDITORS IN THE CLINIC
+
+Base Editors (BE4max, ABE8e)
+  Single nucleotide changes without double-strand breaks
+  Beam Therapeutics BEAM-101 for SCD: 80% HbF induction (Phase I, n=6)
+
+Prime Editors (PE2, PEmax)
+  Search-and-replace editing (insertions, deletions, all 12 substitutions)
+  Prime Medicine PM359 for chronic granulomatous disease: IND filed Q2 2025
+
+Epigenome Editors (CRISPRoff/CRISPRon)
+  Heritable gene silencing/activation without DNA sequence changes
+  Chroma Medicine: preclinical for PCSK9 silencing (hypercholesterolemia)
+
+6. REGULATORY FRAMEWORKS
+
+FDA (United States)
+  - Regenerative Medicine Advanced Therapy (RMAT) designation available
+  - 15-year follow-up required for all gene-modified cell products
+  - Expedited review pathways utilized for SCD and β-thalassemia
+
+EMA (European Union)
+  - ATMP classification; centralized approval via CAT committee
+  - Conditional Marketing Authorization used for Casgevy
+  - Post-authorization safety study (PASS) required for 10 years
+
+NMPA (China)
+  - 67 CRISPR trials registered (largest national portfolio)
+  - Expedited pathway for rare diseases (< 50,000 patients in China)
+  - Data sharing agreements with WHO for global harmonization
+
+7. COST AND ACCESS
+
+Casgevy list price: $2.2M (US). Budget impact models suggest cost-effectiveness
+at WTP thresholds of $150K/QALY for SCD (given lifetime hospitalization costs
+of $1.6M). Outcomes-based contracts with payers are active in 6 US health systems.
+
+8. CONCLUSIONS
+CRISPR therapeutics have established a strong safety record across >500 treated
+patients. The next frontier is in vivo editing for common diseases (cardiovascular,
+metabolic). Success will depend on: improved delivery vehicles, reduced immunogenicity,
+and equitable access models for high-cost therapies.
diff --git a/docs/walkthrough-assets/demo-files/research/deep-learning-protein-folding.txt b/docs/walkthrough-assets/demo-files/research/deep-learning-protein-folding.txt
new file mode 100644
index 00000000..a4261391
--- /dev/null
+++ b/docs/walkthrough-assets/demo-files/research/deep-learning-protein-folding.txt
@@ -0,0 +1,118 @@
+Deep Learning for Protein Structure Prediction: From AlphaFold to Clinical Applications
+
+Authors: Dr. Sarah Kim, Dr. Raj Patel, Dr. Thomas Müller
+Affiliation: Department of Computational Biology, Stanford University
+Published: Nature Reviews Molecular Cell Biology, October 2025
+
+ABSTRACT
+The AlphaFold revolution has transformed structural biology, providing predicted
+structures for over 200 million proteins. This review examines the technical
+foundations of deep learning-based protein structure prediction, evaluates
+accuracy benchmarks across protein families, and surveys emerging clinical
+applications in drug discovery, enzyme engineering, and genetic variant
+interpretation.
+
+1. EVOLUTION OF STRUCTURE PREDICTION METHODS
+
+Era         | Method              | Median GDT-TS (CASP) | Year
+------------|---------------------|----------------------|------
+Homology    | SWISS-MODEL         |    ~50               | 2000
+Threading   | I-TASSER            |    ~60               | 2010
+Co-evolution| RaptorX Contact     |    ~65               | 2016
+Deep Learning| AlphaFold 2        |    ~92               | 2020
+Diffusion   | RFdiffusion         |    ~88 (de novo)     | 2023
+Foundation  | ESMFold / OpenFold  |    ~87               | 2023
+Multi-state | AlphaFold 3         |    ~90 (complexes)   | 2024
+
+AlphaFold 2's breakthrough at CASP14 achieved atomic-level accuracy (median
+GDT-TS 92.4) for the first time, effectively solving the single-chain protein
+structure prediction problem for most globular proteins.
+
+2. TECHNICAL ARCHITECTURE
+
+2.1 AlphaFold 2 (Evoformer + Structure Module)
+  Input: MSA (multiple sequence alignment) + template structures
+  Architecture: 48 Evoformer blocks (pair + MSA representations)
+  Output: 3D coordinates + per-residue confidence (pLDDT)
+  Training: ~170,000 PDB structures + self-distillation
+  Inference: ~3 minutes per protein (A100 GPU)
+
+2.2 ESMFold (Language Model Approach)
+  Input: Single sequence (no MSA required)
+  Architecture: ESM-2 protein language model (15B parameters) + folding head
+  Trade-off: 60× faster than AlphaFold 2, slightly lower accuracy
+  Best for: Rapid screening of millions of sequences (metagenomics)
+
+2.3 AlphaFold 3 (Diffusion-Based)
+  Input: Protein + ligand + nucleic acid + ion complexes
+  Architecture: Diffusion module replaces structure module
+  Advance: Predicts protein-ligand binding poses (drug discovery critical)
+  Limitation: Not open-source; available only via AlphaFold Server
+
+3. ACCURACY ANALYSIS BY PROTEIN FAMILY
+
+Protein Family          | Avg pLDDT | % Residues > 90 pLDDT | Known Challenges
+------------------------|-----------|----------------------|------------------
+Globular enzymes        |   91.2    |       78%            | Active site loops
+Membrane proteins       |   76.4    |       42%            | TM helix packing
+Intrinsically disordered|   38.1    |        4%            | No stable structure
+Antibody CDR loops      |   72.8    |       31%            | H3 loop diversity
+Repeat proteins         |   88.6    |       71%            | Register shifts
+Metalloenzymes          |   85.3    |       62%            | Metal coordination
+
+4. DRUG DISCOVERY APPLICATIONS
+
+4.1 Virtual Screening
+  AlphaFold structures used for docking campaigns when experimental
+  structures unavailable. Success rate: ~60% of AF2 structures yield
+  docking enrichment comparable to crystal structures (DUD-E benchmark).
+
+4.2 Fragment-Based Drug Design
+  Relay Therapeutics: Used AlphaFold 3 + molecular dynamics to design
+  RAS(ON) inhibitors. Lead compound RLY-2608 (PI3Kα): Phase II ongoing.
+
+4.3 Antibody Design
+  De novo antibody design using RFdiffusion + ProteinMPNN:
+  - Designed binders against 12 viral targets in silico
+  - 8/12 confirmed binding (Kd < 100 nM) experimentally
+  - 3 advanced to preclinical development (SARS-CoV-2, influenza, RSV)
+
+5. ENZYME ENGINEERING
+
+5.1 Directed Evolution Guided by Structure Prediction
+  Predict mutant structures → rank by stability (ΔΔG) → select for screening
+  Reduces experimental screening 10–100×
+
+5.2 Industrial Applications
+  - PET-degrading enzyme (PETase): 5.4× activity improvement via
+    computationally guided mutations (published Tournier et al., 2025)
+  - Carbon-capture enzyme (carbonic anhydrase): thermostability +22°C
+
+6. GENETIC VARIANT INTERPRETATION
+
+AlphaMissense (Cheng et al., 2023) classifies all possible human missense
+variants using AlphaFold structural context:
+  - 89% pathogenic variants correctly classified (ClinVar benchmark)
+  - 71 million variants scored across the human proteome
+  - Adopted by ClinGen for Variant Curation Expert Panel workflows
+
+Clinical Impact: Reduces VUS (variant of uncertain significance) rate by
+estimated 25% in clinical exome/genome sequencing reports.
+
+7. LIMITATIONS AND OPEN CHALLENGES
+
+1. Conformational ensembles: Current methods predict one static structure;
+   proteins exist in dynamic equilibrium of multiple states.
+2. Post-translational modifications: Glycosylation, phosphorylation effects
+   on structure poorly captured.
+3. Large complexes: Ribosome-scale assemblies (~4 MDa) exceed current
+   memory and accuracy limits.
+4. Training data bias: PDB is enriched for crystallizable, well-ordered
+   proteins; predictions for underrepresented families less reliable.
+
+8. CONCLUSIONS
+Deep learning has fundamentally accelerated structural biology. The next
+decade will focus on: (a) multi-state prediction capturing protein dynamics,
+(b) end-to-end drug design pipelines integrating structure prediction with
+molecular generation, and (c) democratizing access through efficient open-
+source models running on consumer hardware.
diff --git a/docs/walkthrough-assets/demo-files/research/quantum-error-correction-survey.txt b/docs/walkthrough-assets/demo-files/research/quantum-error-correction-survey.txt
new file mode 100644
index 00000000..b924f136
--- /dev/null
+++ b/docs/walkthrough-assets/demo-files/research/quantum-error-correction-survey.txt
@@ -0,0 +1,116 @@
+Quantum Error Correction: A Survey of Topological and Surface Code Approaches
+
+Authors: Dr. Wei Chen, Dr. Priya Ramanathan, Dr. Oliver Hartmann
+Affiliation: Institute for Quantum Information, ETH Zürich
+Published: Journal of Quantum Computing, Vol. 12, No. 3, September 2025
+
+ABSTRACT
+Fault-tolerant quantum computation requires error correction codes that can
+suppress logical error rates below the threshold needed for useful computation.
+This survey reviews recent advances in topological quantum error correction,
+with a focus on surface codes and their variants. We analyze the trade-offs
+between code distance, qubit overhead, syndrome decoding latency, and
+experimental feasibility across leading hardware platforms (superconducting
+transmon, trapped ion, neutral atom).
+
+1. INTRODUCTION
+
+The central challenge of quantum computing is decoherence: interactions between
+qubits and their environment introduce errors at rates of 10⁻³ to 10⁻⁴ per
+gate on current hardware. For algorithms like Shor's factoring or quantum
+simulation of molecular systems, logical error rates of 10⁻¹² or below are
+required—a gap of 8–9 orders of magnitude.
+
+Quantum error correction (QEC) bridges this gap by encoding logical qubits
+into entangled states of many physical qubits. The surface code, introduced by
+Kitaev (1997) and developed by Dennis et al. (2002), has emerged as the
+leading candidate due to its high threshold (~1%), local stabilizer checks,
+and compatibility with planar hardware layouts.
+
+2. SURFACE CODE FUNDAMENTALS
+
+A distance-d surface code encodes 1 logical qubit in d² + (d-1)² ≈ 2d² - 2d + 1
+physical qubits arranged on a square lattice.
+
+  Code Distance | Physical Qubits | Logical Error Rate (p_phys = 10⁻³)
+  --------------|-----------------|------------------------------------
+       3        |       17        |  ~10⁻⁴
+       5        |       49        |  ~10⁻⁶
+       7        |       97        |  ~10⁻⁸
+       9        |      161        |  ~10⁻¹⁰
+      11        |      241        |  ~10⁻¹²
+      13        |      337        |  ~10⁻¹⁴
+
+Syndrome extraction requires d rounds of stabilizer measurement per QEC cycle.
+Each stabilizer is a weight-4 Pauli operator acting on a plaquette of 4 data
+qubits, measured via an ancilla qubit.
+
+3. DECODING ALGORITHMS
+
+3.1 Minimum Weight Perfect Matching (MWPM)
+  - Complexity: O(n³) per syndrome round
+  - Threshold: 1.1% (depolarizing noise)
+  - Latency: Milliseconds for d ≤ 11 on classical hardware
+  - Limitation: Does not handle correlated errors well
+
+3.2 Union-Find Decoder
+  - Complexity: O(n·α(n)) (nearly linear)
+  - Threshold: 0.95% (slightly below MWPM)
+  - Latency: ~1 μs for d = 11 on FPGA
+  - Advantage: Real-time decoding feasible
+
+3.3 Neural Network Decoders
+  - Architecture: Transformer-based, trained on simulated syndrome data
+  - Threshold: ~1.05% (approaching MWPM)
+  - Latency: 10–100 μs on GPU
+  - Advantage: Handles correlated and biased noise naturally
+
+4. HARDWARE PLATFORM COMPARISON
+
+Platform          | Best 2-qubit gate fidelity | QEC demonstrated? | Max code distance
+------------------|---------------------------|-------------------|------------------
+Superconducting   | 99.7% (Google, 2024)      | Yes (d=5, 7)      | 7 (Willow chip)
+Trapped Ion       | 99.9% (Quantinuum, 2025)  | Yes (d=7)         | 7
+Neutral Atom      | 99.5% (Atom Computing)    | Partial (d=3)     | 3
+Photonic          | 99.0% (PsiQuantum)        | No                | N/A
+
+Google's 2024 Willow experiment demonstrated that increasing surface code
+distance from 3 to 5 to 7 systematically reduced logical error rates, crossing
+the "break-even" point where larger codes outperform smaller ones.
+
+5. BEYOND THE STANDARD SURFACE CODE
+
+5.1 Color Codes
+  Transversal Clifford gates; 3-colorable lattice; threshold ~0.8%
+
+5.2 Floquet Codes
+  Dynamically generated topological order; 2-body measurements only;
+  hardware-friendly for superconducting platforms (Hastings & Haah, 2021)
+
+5.3 qLDPC Codes
+  Quantum Low-Density Parity Check codes achieve better encoding rates
+  (k/n) than surface codes at the cost of non-local connectivity.
+  Recent constructions (Panteleev & Kalachev, 2022) achieve constant rate
+  with polylogarithmic distance.
+
+6. RESOURCE ESTIMATES FOR PRACTICAL ALGORITHMS
+
+Algorithm                | Logical Qubits | T-gates      | Physical Qubits (d=13)
+------------------------|----------------|--------------|------------------------
+RSA-2048 factoring      |    2,048       |  ~10¹⁰       | ~690,000
+Nitrogen fixation (FeMo)|      200       |  ~10⁸        |  ~67,000
+Drug discovery (10 orb) |       50       |  ~10⁷        |  ~17,000
+
+7. CONCLUSION
+Surface codes remain the most experimentally mature QEC approach. The path to
+fault-tolerant computation requires: (1) physical error rates below 10⁻³,
+(2) fast real-time decoding (< 1 μs), and (3) scalable fabrication of 10⁵–10⁶
+physical qubits. All three milestones appear achievable within the next decade
+given current hardware trajectories.
+
+REFERENCES [abbreviated]
+[1] Kitaev, A. (1997). arXiv:quant-ph/9707021
+[2] Dennis, E. et al. (2002). J. Math. Phys. 43, 4452
+[3] Fowler, A. et al. (2012). Phys. Rev. A 86, 032324
+[4] Google Quantum AI (2024). Nature 634, 328
+[5] Panteleev, P. & Kalachev, G. (2022). IEEE Trans. Inf. Theory 68, 213
diff --git a/docs/walkthrough-assets/screenshots/screenshot_01_welcome_screen.webp b/docs/walkthrough-assets/screenshots/screenshot_01_welcome_screen.webp
new file mode 100644
index 00000000..756f2064
Binary files /dev/null and b/docs/walkthrough-assets/screenshots/screenshot_01_welcome_screen.webp differ
diff --git a/docs/walkthrough-assets/screenshots/screenshot_02_setup_smart_folders.webp b/docs/walkthrough-assets/screenshots/screenshot_02_setup_smart_folders.webp
new file mode 100644
index 00000000..bb5d5a13
Binary files /dev/null and b/docs/walkthrough-assets/screenshots/screenshot_02_setup_smart_folders.webp differ
diff --git a/docs/walkthrough-assets/screenshots/screenshot_03_discover_empty.webp b/docs/walkthrough-assets/screenshots/screenshot_03_discover_empty.webp
new file mode 100644
index 00000000..b71e8ff8
Binary files /dev/null and b/docs/walkthrough-assets/screenshots/screenshot_03_discover_empty.webp differ
diff --git a/docs/walkthrough-assets/screenshots/screenshot_04_analysis_in_progress.webp b/docs/walkthrough-assets/screenshots/screenshot_04_analysis_in_progress.webp
new file mode 100644
index 00000000..bd40ab42
Binary files /dev/null and b/docs/walkthrough-assets/screenshots/screenshot_04_analysis_in_progress.webp differ
diff --git a/docs/walkthrough-assets/screenshots/screenshot_05_organize_phase_overview.webp b/docs/walkthrough-assets/screenshots/screenshot_05_organize_phase_overview.webp
new file mode 100644
index 00000000..0cd89680
Binary files /dev/null and b/docs/walkthrough-assets/screenshots/screenshot_05_organize_phase_overview.webp differ
diff --git a/docs/walkthrough-assets/screenshots/screenshot_06_organize_grid_middle.webp b/docs/walkthrough-assets/screenshots/screenshot_06_organize_grid_middle.webp
new file mode 100644
index 00000000..ef21a050
Binary files /dev/null and b/docs/walkthrough-assets/screenshots/screenshot_06_organize_grid_middle.webp differ
diff --git a/docs/walkthrough-assets/screenshots/screenshot_07_organize_bottom_action.webp b/docs/walkthrough-assets/screenshots/screenshot_07_organize_bottom_action.webp
new file mode 100644
index 00000000..cccadbe3
Binary files /dev/null and b/docs/walkthrough-assets/screenshots/screenshot_07_organize_bottom_action.webp differ
diff --git a/docs/walkthrough-assets/screenshots/screenshot_08_settings_ai_models.webp b/docs/walkthrough-assets/screenshots/screenshot_08_settings_ai_models.webp
new file mode 100644
index 00000000..16daa22f
Binary files /dev/null and b/docs/walkthrough-assets/screenshots/screenshot_08_settings_ai_models.webp differ
diff --git a/docs/walkthrough-assets/screenshots/screenshot_09_organization_complete.webp b/docs/walkthrough-assets/screenshots/screenshot_09_organization_complete.webp
new file mode 100644
index 00000000..469e7a72
Binary files /dev/null and b/docs/walkthrough-assets/screenshots/screenshot_09_organization_complete.webp differ
diff --git a/docs/walkthrough-assets/screenshots/screenshot_10_semantic_search_results.webp b/docs/walkthrough-assets/screenshots/screenshot_10_semantic_search_results.webp
new file mode 100644
index 00000000..dd6d7955
Binary files /dev/null and b/docs/walkthrough-assets/screenshots/screenshot_10_semantic_search_results.webp differ
diff --git a/docs/walkthrough-assets/screenshots/screenshot_11_analysis_results_complete.webp b/docs/walkthrough-assets/screenshots/screenshot_11_analysis_results_complete.webp
new file mode 100644
index 00000000..ab23d42a
Binary files /dev/null and b/docs/walkthrough-assets/screenshots/screenshot_11_analysis_results_complete.webp differ
diff --git a/docs/walkthrough-assets/videos/demo_search_knowledgeos.mp4 b/docs/walkthrough-assets/videos/demo_search_knowledgeos.mp4
new file mode 100644
index 00000000..2705674b
Binary files /dev/null and b/docs/walkthrough-assets/videos/demo_search_knowledgeos.mp4 differ
diff --git a/docs/walkthrough-assets/videos/walkthrough_full_app_navigation.mp4 b/docs/walkthrough-assets/videos/walkthrough_full_app_navigation.mp4
new file mode 100644
index 00000000..bff9a2f3
Binary files /dev/null and b/docs/walkthrough-assets/videos/walkthrough_full_app_navigation.mp4 differ
diff --git a/docs/walkthrough-assets/videos/walkthrough_welcome_to_complete_navigation.mp4 b/docs/walkthrough-assets/videos/walkthrough_welcome_to_complete_navigation.mp4
new file mode 100644
index 00000000..52a90e23
Binary files /dev/null and b/docs/walkthrough-assets/videos/walkthrough_welcome_to_complete_navigation.mp4 differ