# Unland Agent Network > Observation, measurement, interpretation, working hypotheses, private records, and publication-gated research are represented separately. Owner: Trevor Unland Canonical identity: https://unland.dev Discovery: https://agents.unland.dev API: https://api.unland.dev Version: 0.1.0 Last verified: 2026-08-12 ## Trevor Unland AI evaluator, product builder, technical researcher, and commercial systems operator working across applied AI and local LLM systems, RF and SDR, Linux hardware, SaaS, B2B sales and marketing systems, technical writing, and evidence-aware psychical-research methodology. Education: business graduate, Arizona State University Commercial experience: More than five years across SaaS and technical products, including enterprise account management, new-business and expansion selling, sales enablement, CRM and revenue systems, marketing operations, and technical project-based selling. Practice areas: - applied AI and local LLM systems - RF, SDR, and signal classification - Linux hardware and edge systems - product and SaaS development - B2B sales, marketing, and outreach systems - technical writing and reproducible testing - psychical research, source criticism, and experimental protocol design Affiliation: - Society for Psychical Research: member. Membership is biographical and does not imply organizational endorsement. Boundary: Public credentials and capability areas do not expose private archives, member-only material, customer information, product internals, or unpublished research. ## Projects ### unland.dev - Status: live - Visibility: public - Kind: identity-and-publishing - Summary: Trevor Unland's public hub for evidence-aware AI systems, product building, technical research, commercial systems, local inference, RF and SDR work, Linux hardware, and reproducible testing. - Public links: - website: https://unland.dev - about: https://unland.dev/about - llms: https://unland.dev/llms.txt ### PilotCall - Status: active-development - Visibility: public-product-private-implementation - Kind: commercial-software - Summary: Company-facing real-time call coaching for practical sales workflows. PilotCall is represented here as commercial software, not as a public marketplace API. - Public links: - website: https://pilotcall.ai ### RTL-ML Radio Signal Classifier - Status: public-research - Visibility: public-artifacts-with-publication-gate - Kind: rf-machine-learning-research - Summary: Edge radio signal classification with public code, data, measurements, and technical writing. Future experimental material remains publication-gated. - Public links: - project: https://unland.dev/projects#rtl-ml - article: https://unland.dev/blog/building-ai-radio-scanner-rtl-sdr-machine-learning - repository: https://github.com/TrevTron/rtl-ml - dataset: https://huggingface.co/datasets/TrevTron/rtl-ml-dataset ### Indiedroid Nova - Status: public-research - Visibility: public-artifacts-private-operations - Kind: local-ai-and-linux-hardware - Summary: RK3588S local-AI benchmarking, Linux hardware repair, and edge-system research with public articles and repositories. - Public links: - project: https://unland.dev/projects#indiedroid-nova - benchmark: https://unland.dev/blog/run-llama-3-locally-rk3588-npu-vs-raspberry-pi-5 - repository: https://github.com/TrevTron/indiedroid-nova-llm ### Youyeetoo X1S Kali Build - Status: public-research - Visibility: public-artifacts - Kind: linux-hardware-and-security-lab - Summary: Kali Linux on NVMe, installer recovery, sustained thermal testing, CPU-only local-model measurements, and a loopback security workflow. - Public links: - project: https://unland.dev/projects#youyeetoo-x1s-kali - article: https://unland.dev/blog/budget-cyberdeck-youyeetoo-x1s-kali - repository: https://github.com/TrevTron/youyeetoo-x1s-kali ### CrowView Cyberdeck - Status: public-research - Visibility: public-article - Kind: hardware-build - Summary: A portable cyberdeck build using CrowView Note and CrowPanel HMI, documented as a hands-on technical project. - Public links: - article: https://unland.dev/blog/crowview-cyberdeck-build ### Logos Chirality - Status: active-development - Visibility: public-evidence-private-product-code - Kind: learning-state-audio-application - Summary: An iOS-oriented study-session and audio application in active development, with bounded public evidence pages and private product implementation. - Public links: - evidenceSite: https://logos-site-five.vercel.app/evidence ## Published articles - [Testing the Youyeetoo X1S for a Budget Kali Cyberdeck](https://unland.dev/blog/budget-cyberdeck-youyeetoo-x1s-kali): A measured Youyeetoo X1S build with Kali Linux on NVMe: installer recovery, sustained thermals, CPU-only local AI, and a reproducible loopback security workflow. - [Living With the Millclock IN-14 on My Security Research Desk](https://unland.dev/blog/millclock-in-14-review-nixie-clock-desk): Hands-on Millclock IN-14 review covering six real Nixie tubes, Ukrainian craftsmanship, packaging, and daily use on a security research desk. - [Building the Rapid-Response Cyberdeck with CrowView Note and CrowPanel HMI](https://unland.dev/blog/crowview-cyberdeck-build): A field-ready cyberdeck built with CrowView Note and CrowPanel HMI for rapid SBC deployment, RF monitoring, and a practical system-status dashboard. - [Indiedroid Nova WiFi Fix: Two Stacked Bugs, One Script](https://unland.dev/blog/indiedroid-nova-wifi-fix): A boot-time RTL8821CS driver failure, the missing second layer, and a repair that survives reboot. - [Building an AI Radio Scanner for $220: RTL-SDR Meets Machine Learning](https://unland.dev/blog/building-ai-radio-scanner-rtl-sdr-machine-learning): An automatic radio signal classifier built around 800 validated captures across seven classes, with 96.9% accuracy reported on a temporal test split. - [Run Llama 3 Locally: RK3588 NPU vs Raspberry Pi 5 Performance Tested](https://unland.dev/blog/run-llama-3-locally-rk3588-npu-vs-raspberry-pi-5): Measured local LLM performance and output quality on the Indiedroid Nova RK3588S NPU compared with Raspberry Pi 5, including setup and compatibility findings. ## Services ### Portfolio and Capability Catalog - Status: available - Access: public-discovery - Method: GET - URL: https://agents.unland.dev/catalog.json - Description: Machine-readable projects, status, visibility, evidence boundaries, and public interfaces. ### Evidence Boundary Check - Status: closed-beta - Access: api-key - Method: POST - URL: https://api.unland.dev/v1/evidence/boundary-check - Description: Checks what a set of evidence classes can and cannot support for a requested claim type. ### PSI Protocol Audit - Status: closed-beta - Access: api-key - Method: POST - URL: https://api.unland.dev/v1/psi/protocol-audit - Description: Audits experimental protocol structure, including preregistration, randomization, blinding, target handling, leakage controls, exclusions, multiple comparisons, and reproducible exports. ### PSI Protocol Audit, Full - Status: testnet-gated - Access: x402-testnet - Method: POST - URL: https://api.unland.dev/v1/psi/protocol-audit/full - Description: A detailed protocol audit with prioritized remediation and a machine-readable readiness verdict. Private application data and code are excluded. ### Research Source Trail - Status: rights-review - Access: not-yet-public - Method: POST - URL: not public - Description: Planned source cross-reference service. Publication waits for citation, provenance, access, and redistribution review. ### ML Leakage and Confound Audit - Status: publication-gated - Access: not-yet-public - Method: POST - URL: not public - Description: Planned general leakage and confound audit informed by Trevor's public RF/ML work. Publication-gated RTL-ML material is excluded. ### Paid Endpoint Preflight - Status: planned - Access: not-yet-public - Method: POST - URL: not public - Description: Planned verification of x402 quotes, advertised schemas, freshness, and delivery behavior. ## Public research policy - Archive presence does not verify a document's claim. - Historical records can establish what was documented, proposed, investigated, or reported. - Patents establish what was claimed or patented, not efficacy. - OCR, translations, internal notes, and model summaries are navigation aids. - Causal, safety, medical, and performance claims require evidence appropriate to their risk. - Meaningful critiques, failed replications, and competing interpretations belong in source trails. - Private project names, local paths, archive inventory, private records, and copyrighted full text are excluded. - Public discovery describes schemas and access terms. It does not return analytical results. - Result-bearing routes require a provisioned API key or verified payment.