` marker. Import it as-is into Claude / Gemini / any agent: paste this file as base knowledge. Links between concepts are bundle-relative (/labs/x.md). --> --- okf_version: 2.0.0 title: OKF bundle — scope public source: askmojo scope: public concept_counts: lab: 1 magik: 2 creator: 1 output: 1 --- # OKF bundle (scope `public`) Open Knowledge Format v2.0.0 bundle — markdown + YAML frontmatter. The AskMojo database stays the source of truth; this bundle is an exported view filtered by scope. ## Concepts - **lab** : 1 - **magik** : 2 - **creator** : 1 - **output** : 1 ## Indexes - [creators](/creators/index.md) - [labs](/labs/index.md) - [outputs](/outputs/index.md) --- title: Creators description: Creators index count: 1 --- # Creators - [Jason Wright](/creators/jason-wright.md) (`public`) --- type: creator title: Jason Wright description: I run sales & marketing ops, and I was sick of my team burning days on research instead of closing. I build labs that find, qualify and reach the right people on autopilot. Copy them, plug in your stack, and let them prospect for you. resource: /creators/jason-wright timestamp: '2026-06-29T10:20:58.959Z' visibility: public --- I run sales & marketing ops, and I was sick of my team burning days on research instead of closing. I build labs that find, qualify and reach the right people on autopilot. Copy them, plug in your stack, and let them prospect for you. --- title: Labs description: Labs index count: 3 --- # Labs - [Research & Reports](/labs/research-reports.md) (`public`) - [Deep research report](/labs/research-reports/magiks/deep-research-report.md) (`public`) - [Competitor benchmark](/labs/research-reports/magiks/competitor-benchmark.md) (`public`) --- type: lab title: Research & Reports description: Turn any topic into deep, sourced research — full research reports on any subject and side-by-side competitor benchmarks, delivered as clean, branded reports. resource: /labs/research-reports timestamp: '2026-06-18T09:45:23.800Z' visibility: public language: en creator: /creators/jason-wright.md --- # Research & Reports Turn any topic into deep, sourced research — full research reports on any subject and side-by-side competitor benchmarks, delivered as clean, branded reports. Creator : [/creators/jason-wright.md](/creators/jason-wright.md) ## Magiks - [Deep research report](/labs/research-reports/magiks/deep-research-report.md) — Turn any topic into deep, sourced research delivered as a clean, branded AskMojo PDF, deep dives, competitor benchmarks, market briefs and exec summaries. - [Competitor benchmark](/labs/research-reports/magiks/competitor-benchmark.md) — Turn any topic into deep, sourced research delivered as a clean, branded AskMojo PDF, deep dives, competitor benchmarks, market briefs and exec summaries. ## Widgets - **Mission** (note, mode=static) - **Latest research reports** (magik-outputs, mode=static) - **Deep research report** (run-magik, mode=static) - **Latest benchmarks** (magik-outputs, mode=static) - **Competitor benchmark** (run-magik, mode=static) --- type: magik title: Competitor benchmark description: Turn any topic into deep, sourced research delivered as a clean, branded AskMojo PDF, deep dives, competitor benchmarks, market briefs and exec summaries. resource: /labs/research-reports?magik=competitor-benchmark tags: - exa - perplexity timestamp: '2026-06-29T08:30:18.401Z' visibility: public language: en lab: /labs/research-reports.md output_type: pdf --- # Competitor benchmark Turn any topic into deep, sourced research delivered as a clean, branded AskMojo PDF, deep dives, competitor benchmarks, market briefs and exec summaries. Lab : [/labs/research-reports.md](/labs/research-reports.md) ## Skill --- name: competitor-benchmark description: Side-by-side competitor benchmark as a branded PDF. user-invocable: true category: research tools: [exa, perplexity] --- # Competitor benchmark Research and compare 3 to 6 competitors side by side: positioning, pricing, features, strengths, and gaps. ## Inputs - Your market or company name and what you do - Competitor names (3 to 6), or "find the main ones in this space" - Focus dimension: pricing, product features, marketing, or full benchmark ## Steps 1. Research each competitor: positioning, pricing tier, key features, and differentiation angle 2. Identify each competitor's clearest strength and most visible weakness 3. Build a comparison table covering all dimensions consistently 4. Write a 1-page synthesis with a clear strategic recommendation and your differentiating angle ## Format Comparison table + 1-page synthesis. Every claim linked to a real source. Output: branded PDF. ## Tone Analytical and direct. Useful for a strategy meeting, not just interesting to read. --- type: magik title: Deep research report description: Turn any topic into deep, sourced research delivered as a clean, branded AskMojo PDF, deep dives, competitor benchmarks, market briefs and exec summaries. resource: /labs/research-reports?magik=deep-research-report tags: - exa - perplexity timestamp: '2026-06-29T08:30:18.401Z' visibility: public language: en lab: /labs/research-reports.md output_type: pdf --- # Deep research report Turn any topic into deep, sourced research delivered as a clean, branded AskMojo PDF, deep dives, competitor benchmarks, market briefs and exec summaries. Lab : [/labs/research-reports.md](/labs/research-reports.md) ## Skill --- name: deep-research-report description: Deep, sourced research on any topic, delivered as a branded PDF. user-invocable: true --- # Deep research report Take the topic the user gives you. Run thorough, multi-source web research (use built-in search). Produce a structured, sourced report: executive summary, key findings (with citations), analysis, risks/opportunities, and a sources list. Format it for a clean, branded AskMojo PDF — clear headings, short paragraphs, bullet takeaways. Never fabricate sources; cite real URLs. --- scope: public --- # Log Historique des changements de visibilité / publication (dates ISO). _Aucun évènement de publication horodaté._ --- type: output title: Deep research report — July 3, 2026 resource: /o/268aeec6-b9f7-4a47-9059-c2191f8bf407 timestamp: '2026-07-03T23:53:34.482Z' visibility: public magik: /labs/research-reports/magiks/deep-research-report.md --- # Deep research report — July 3, 2026 Généré par : [/labs/research-reports/magiks/deep-research-report.md](/labs/research-reports/magiks/deep-research-report.md) # Owning Your AI in 2026: The Portable Workspace Revolution The age of locked-in chatbots is ending. Creators, developers, and businesses are moving to portable, exportable AI workspaces where they own their data, control their tools, and decide their own exit terms. This shift mirrors the broader creator economy trend: from renting audience on platforms to owning distribution. July 2026 marks an inflection point—regulatory pressure, open-source alternatives, and visible platform failures are converging to make AI ownership the default, not the exception. --- ## 🔒 The Lock-In Problem: What You Lose When Platforms Own Your AI **ChatGPT Business workspaces have no export button** *Users who merge personal data into Business plans cannot retrieve it—ever* OpenAI's ChatGPT Business (formerly Team) workspaces create permanent vendor lock-in. Once a user merges their personal workspace into a Business plan, that data becomes irretrievable. There is no export function, no automated download, and no machine-readable format available. Users who leave the workspace or whose workspace is deactivated lose access to all conversation history, prompts, and context. OpenAI's official help documentation confirms this is not a bug—it's by design. **The GDPR violation hiding in plain sight** *EU law guarantees data retrieval rights; platform terms often ignore them* Under GDPR Articles 15 and 20, users have enforceable rights to access and port their data. ChatGPT Business's lack of an export option is a documented violation, with users forced to submit formal privacy requests through OpenAI's Privacy Portal—a process that does not deliver structured, reusable data. Community forums describe support suggesting manual copy-pasting as the only workaround, a solution deemed "unprofessional" for business-scale operations. **The upsell trap: export only for Enterprise** *Core portability features locked behind significantly higher pricing* OpenAI offers rich export tools—prompts, files, outputs, conversation history—but only to Enterprise customers paying substantially more than Business-tier users. This pricing structure effectively holds user data hostage, forcing businesses to either upgrade or remain locked in indefinitely. > In 2026, if your AI knowledge lives inside one platform, you don't own it. You're renting. And the landlord can change the locks. --- ## 📊 The Creator Economy Parallel: From Algorithm Dependence to Owned Audiences **56.5% of the $252 billion creator economy is now independent** *Revenue captured by solo creators, not media networks—a decisive ownership shift* The 2026 creator economy, valued at $252 billion, shows a structural shift: 56.5% of revenue now flows to independent creators rather than large platforms or networks. This mirrors what's happening in AI tooling—users are moving from platform-dependent workflows to owned infrastructure. Just as creators discovered that owned email lists and communities offer more stability than algorithmic reach, AI users are learning that self-hosted workspaces outlast vendor roadmaps. **Platform independence as revenue stability** *88% of community builders use membership models; ownership is the business model* - **48%** of creators now run entirely solo businesses, functioning as solopreneurs - **88%** of community builders monetize through membership models they control - **18,000+** active communities operate on owned platforms like Circle, controlling access and member relationships - **58%** of creators maintain presences on 3+ platforms to mitigate algorithm risk - **45.2%** of revenue still tied to social platforms, driving demand for platform-agnostic tools The lesson is clear: diversification and ownership reduce risk. First-party data—audience contacts, conversation history, custom prompts—is now treated as a balance-sheet asset, not a ephemeral resource trapped in someone else's database. --- ## 🚀 The Self-Hosted Awakening: Odysseus and the MIT-Licensed Alternative **PewDiePie releases Odysseus in June 2026** *22,400 GitHub stars in weeks; a local-first, privacy-first AI workspace for everyone* In June 2026, Felix Kjellberg (PewDiePie) released Odysseus, an open-source, MIT-licensed AI workspace that runs entirely on user hardware. It is not a new AI model—it's a complete command center consolidating chat, autonomous agents, deep research, coding, email, calendar, and persistent memory into a single, self-hosted interface. Within weeks, it reached 22,400 stars and 2,800 forks on GitHub, signaling mainstream appetite for platform independence. **Local-first architecture: your laptop, your data, your rules** *All computation happens on infrastructure you control; no third-party data leakage* - **Model agnostic**: integrates with Ollama (for local LLaMA, Mistral, Qwen models) and cloud APIs (OpenAI, Anthropic, OpenRouter) - **AI agents**: tool use via MCP (Model Context Protocol), web search, shell access, file operations - **Deep research**: adapted from Tongyi DeepResearch for multi-step analysis workflows - **Email & calendar**: AI triage, summaries, auto-replies, CalDAV sync - **Code editor**: AI-assisted coding with syntax highlighting - **Hardware-aware**: includes 270+ model recommendations based on your GPU/CPU Installation is straightforward: ```bash git clone https://github.com/pewdiepie-archdaemon/odysseus.git cd odysseus && docker compose up -d --build ``` Access via `http://localhost:7000`. No cloud dependency, no subscription, no data leaving your network unless you explicitly route to external APIs. **Cost-effective and production-ready** *$0 software cost; competes directly with ChatGPT, Claude, and Codex* Odysseus represents a shift away from monthly SaaS subscriptions and corporate data capture. Users gain full control at zero software cost, paying only for cloud API calls if they choose to use proprietary models. For teams with access to local GPUs, the entire stack can run air-gapped. --- ## 🛠️ The Emerging Ecosystem: Tools Built for Portability **Kept: local-first AI conversation archive** *Search, archive, and recall conversations from ChatGPT, Claude, Gemini, Grok, Kimi* Released in May 2026 with 107 GitHub stars, Kept is a browser extension and local app that archives AI conversations from multiple platforms into a searchable, local-first knowledge graph. MIT-licensed, it gives users a unified memory layer across all their AI tools, eliminating platform silos. **TransferLLM: migrate chat history between AI platforms** *10,000+ migrations completed; runs locally with zero server reads* TransferLLM enables private, automated migration of conversation history between AI tools. Instead of CSV exports and manual reformatting, it connects to both source and destination platforms via OAuth or secure session credentials and rebuilds conversation threads natively. The entire process runs locally, ensuring no third-party ever reads prompts or outputs. **MemoryLake and CloudFuze: migration guides for Gemini, Claude, Copilot** *Practical tooling for teams switching between enterprise AI platforms* - **MemoryLake** published a complete guide in May 2026 for migrating Gemini memory (Gems, Saved Info, Workspace context) into Claude Projects - **CloudFuze** offers Gemini-to-Copilot migration, converting conversations into OneNote files stored in OneDrive - **Deplexity** (released May 2026): CLI tool to export Perplexity AI conversations, spaces, and profiles These tools reflect real-world demand: teams want to move between platforms without losing institutional knowledge. **Synapse: self-hosted AI workspace with shared teammates** *Shareable AI agents, shared conversations, memory, governed access to MCP tools* Synapse (62 stars, Apache 2.0 license) is a self-hosted workspace enabling teams to deploy AI teammates with persistent memory, shared conversations, and governed access to plugins, MCP tools, and local devices. It's designed for organizations that need collaborative AI without cloud dependency. --- ## ⚖️ Regulatory Tailwinds: The EU Data Act and Mandatory Portability **Core provisions in effect since September 2025** *Users gain enforceable rights to access and port data they generate* The EU Data Act's core portability provisions began applying on 12 September 2025. Users now have enforceable rights to access and port data they generate, regardless of platform. Providers must remove technical and contractual barriers to switching, ensuring customers can terminate cloud/SaaS/PaaS/IaaS contracts with a maximum two-month notice and port data within 30 days. **Switching fees eliminated from January 2027** *All data transfer and reformatting charges prohibited; only limited multicloud exceptions* Starting 12 January 2027, all switching charges—including data egress fees—are entirely prohibited. Between now and then, providers may only charge cost-covering fees (pass-through basis, no markup). This removes the financial penalty that historically locked enterprises into cloud vendors. **Functional equivalence required for transitions** *IaaS providers must ensure documentation, capabilities, and support during migrations* - **30-day portability window**: customers must be able to port data within 30 days of notice - **Free, open interfaces**: providers must offer accessible APIs and data formats - **Harmonised standards**: the Act mandates open interoperability specifications across sectors, including AI - **GDPR complement**: builds on GDPR Article 20 by imposing vendor-specific obligations for machine-readable formats Non-compliance results in fines set at Member State levels. The Act applies to all "data processing services," which includes AI platforms that store user prompts, outputs, and context. --- ## 🎯 What Creators and Businesses Gain from Portable AI **Full data sovereignty** *You decide where data lives, who can access it, and when to delete it* Self-hosted and exportable AI workspaces give users complete control over their data. Sensitive prompts, proprietary research, client information, and institutional knowledge stay within your infrastructure. You set retention policies, govern access, and ensure compliance without relying on a vendor's security promises. **Vendor agnosticism and risk mitigation** *Switch models, providers, or entire platforms without losing conversation history or context* - **Model flexibility**: test GPT-4, Claude 3.5, Gemini 1.5, or local Mistral/LLaMA models within the same workspace - **Exit strategy**: if a vendor raises prices, shuts down a feature, or violates trust, you can migrate instantly - **No hostage data**: conversation history, custom instructions, and memory persist across platform changes - **Regulatory resilience**: when governments restrict access to certain models (e.g., Anthropic's Fable 5 US government block), self-hosted users continue unaffected **Customization and workflow integration** *Build agentic AI tailored to your unique processes, not generic SaaS features* Portable workspaces enable deep customization: integrate your own tools, plugins, and APIs; build workspace-specific agents for grants, content pipelines, or customer support; automate end-to-end workflows that reflect how your business actually operates. You're no longer limited by a vendor's feature roadmap. **Cost control at scale** *Avoid per-seat SaaS pricing; pay only for compute you use* Self-hosted AI eliminates recurring per-user fees. Teams with existing GPU infrastructure or access to cloud compute can run models at marginal cost, scaling usage without scaling subscription costs. For high-volume users, this translates to significant savings. **Privacy by design** *Local-first architecture means no data ever leaves your network unless you explicitly choose* Open-source workspaces like Odysseus and Synapse route all computation locally by default. Sensitive data never touches a vendor's servers unless you explicitly configure external API calls. For legal, healthcare, finance, and government teams, this is not a nice-to-have—it's a requirement. **Institutional memory that survives vendor changes** *Knowledge graphs, conversation archives, and custom agents persist regardless of platform churn* Tools like Kept and TransferLLM enable users to build a unified, portable memory layer across all AI interactions. Your organization's collective AI knowledge becomes a durable asset, not a collection of siloed chat logs locked in vendor databases. --- ## 🌐 The Shift Is Structural, Not Speculative **Google shut down an open-source tool with two days' notice** *Vendor risk is real; self-hosted stacks survive platform decisions* In June 2026, Google shut down a widely used open-source tool with only 48 hours' notice. Users who built daily workflows around it faced sudden disruption. At the same time, Anthropic disabled Fable 5 due to US government restrictions. Dan Cucolea, creator of the Hermes Stack, noted: "My setup doesn't care, and here's why yours shouldn't either." Self-hosted infrastructure decouples your operations from vendor roadmaps and policy changes. **Platform account suspensions with no human review** *AI creators losing Instagram, YouTube access with no appeal—your AI workspace can be next* In May 2026, an AI creator's Instagram account was disabled with the message: "We've reviewed your account and found that it still doesn't follow our Community Standards on account integrity." No explanation. No human review. No appeal. The same risk applies to AI platforms: terms of service violations, automated moderation errors, or policy shifts can lock you out of your own data instantly. Portability is insurance. **Third-party AI adoption growing 200% since 2023** *Google Workspace users installing third-party AI tools alongside native Gemini* AI add-on installs in Google Workspace have grown over 200% since 2023, with users adopting third-party tools (Perplexity, Claude) in parallel with native AI (Gemini). This signals demand for platform-agnostic solutions that work across ecosystems, not lock users into a single vendor's stack. **The $1 trillion creator economy by 2032** *Ownership-driven models are the foundation of the next phase of digital work* The creator economy is projected to exceed $1 trillion by 2032, driven by convergence of commerce and content, live-shopping, affiliate integrations, and owned audiences. The same dynamics are reshaping AI tooling: users demand portability, interoperability, and data sovereignty. The platforms that win will be those that enable ownership, not those that lock it away. --- ## 📚 Sources https://allagentsconsidered.substack.com/p/i-built-my-ai-stack-to-survive-vendor https://penfieldlabs.substack.com/p/the-platform-changes-the-game-is https://theworldtrendnow.com/creator-economy-2026-platform-independence-over-algorithm-dependence/ https://eimoh.com/creators-ditch-platforms-for-private-fiefdoms-built-on-email-and-apis/ https://github.com/zai-org/Synapse https://github.com/egroup-labs/kept https://github.com/clappingmonkey/Deplexity https://github.com/pewdiepie-archdaemon/odysseus https://aibusinesslab.ai/building-an-ai-second-brain-you-own/ https://www.mindstudio.ai/blog/self-hosted-ai-workspaces-vs-cloud-platforms-2 https://blog.corenexis.com/self-hosted-ai-odysseus https://medium.com/@sidhanthpandey/every-ai-knows-you-now-none-of-it-is-yours-14b0c42d4790 https://dev.to/agentrisk/you-dont-own-your-ai-agent-and-even-if-you-did-would-you-trust-it-4d2m https://www.memorylake.ai/en/blogs/migrate-gemini-memory-to-claude-projects https://www.cloudfuze.com/gemini-to-copilot-migration/ https://transferllm.com/ https://community.openai.com/t/urgent-gdpr-violation-no-export-option-for-chatgpt-team-workspace-data/1090657 https://www.linkedin.com/pulse/twenty-2026-creator-economy-predictions-lyle-stevens-kalgc https://circle.so/blog/creator-economy-statistics https://explainx.ai/blog/odysseus-self-hosted-ai-workspace-2026 https://digital-strategy.ec.europa.eu/en/factpages/data-act-explained https://www.lw.com/en/insights/eu-data-act-significant-new-switching-requirements-due-to-take-effect-for-data-processing-services --- title: Outputs description: Outputs index count: 1 --- # Outputs - [Deep research report — July 3, 2026](/outputs/268aeec6-b9f7-4a47-9059-c2191f8bf407.md) (`public`)