THE AI AGENT FIELD GUIDESOURCES FIRST. CLEARER CHOICES.
THE EDITORIAL RECORD

What changed in the directory.

This initial record documents sourced profile publication. Future entries will describe material fact changes. Website rebuilds do not update source dates.

Shopify Sidekick →

Initial sourced profile published. Documented vendor claims separated from unverified fields.

1 reviewed sources · initial profile, capabilities, pricing, source review

Helium Agent →

Initial sourced profile published. Documented vendor claims separated from unverified fields.

1 reviewed sources · initial profile, capabilities, pricing, source review

Kalodata →

Initial sourced profile published. Documented vendor claims separated from unverified fields.

1 reviewed sources · initial profile, capabilities, pricing, source review

Gorgias AI Agent →

Initial sourced profile published. Documented vendor claims separated from unverified fields.

2 reviewed sources · initial profile, capabilities, pricing, source review

Tidio Lyro →

Initial sourced profile published. Documented vendor claims separated from unverified fields.

1 reviewed sources · initial profile, capabilities, pricing, source review

Photoroom →

Initial sourced profile published. Documented vendor claims separated from unverified fields.

2 reviewed sources · initial profile, capabilities, pricing, source review

Creatify Agent →

Initial sourced profile published. Documented vendor claims separated from unverified fields.

2 reviewed sources · initial profile, capabilities, pricing, source review

OpusClip →

Initial sourced profile published. Documented vendor claims separated from unverified fields.

2 reviewed sources · initial profile, capabilities, pricing, source review

Descript →

Initial sourced profile published. Documented vendor claims separated from unverified fields.

2 reviewed sources · initial profile, capabilities, pricing, source review

ElevenLabs →

Initial sourced profile published. Documented vendor claims separated from unverified fields.

2 reviewed sources · initial profile, capabilities, pricing, source review

Jasper →

Initial sourced profile published. Documented vendor claims separated from unverified fields.

1 reviewed sources · initial profile, capabilities, pricing, source review

n8n →

Initial sourced profile published. Documented vendor claims separated from unverified fields.

2 reviewed sources · initial profile, capabilities, pricing, source review

Zapier Agents →

Initial sourced profile published. Documented vendor claims separated from unverified fields.

1 reviewed sources · initial profile, capabilities, pricing, source review

Elicit →

Initial sourced profile published. Documented vendor claims separated from unverified fields.

2 reviewed sources · initial profile, capabilities, pricing, source review

Harvey →

Initial sourced profile published. Documented vendor claims separated from unverified fields.

1 reviewed sources · initial profile, capabilities, pricing, source review

Patsnap Eureka →

Initial sourced profile published. Documented vendor claims separated from unverified fields.

1 reviewed sources · initial profile, capabilities, pricing, source review

Plantix →

Initial sourced profile published. Documented vendor claims separated from unverified fields.

1 reviewed sources · initial profile, capabilities, pricing, source review

AlphaSense →

Initial sourced profile published. Documented vendor claims separated from unverified fields.

1 reviewed sources · initial profile, capabilities, pricing, source review

Nansen →

Initial sourced profile published. Documented vendor claims separated from unverified fields.

1 reviewed sources · initial profile, capabilities, pricing, source review

Chainalysis →

Initial sourced profile published. Documented vendor claims separated from unverified fields.

1 reviewed sources · initial profile, capabilities, pricing, source review

Tidio Lyro →

Recorded three first replies in the official Lyro Patagonia public demo after the earlier loading blockage. All 11 frozen visible conditions passed; separate official-policy review retains unsupported dealer/contact assumptions and omitted order-number and other-proof qualifications. The limited public feature record does not execute either original Tidio product case or establish policy accuracy or overall product quality.

1 reviewed sources · public feature test, first replies, frozen visible-condition review, policy fact-check limitations

Google Patents →

Executed both frozen identifier cases once: exact retrieval partial (4/5 conditions passed; current legal status unverified), malformed identifier rejection passed 4/4. Retained first states and original PDF; earlier source reads remain separate.

1 reviewed sources · native search evaluation, original publication evidence, malformed identifier validation

Rebuy →

Profile added from fetched official sources. Vendor statements, proposed evaluation, unknown commercial terms, and unperformed live tests are distinguished.

2 reviewed sources · initial profile, capabilities, pricing mode, setup guidance, source review

HeyGen →

Profile added from fetched official sources. Vendor statements, proposed evaluation, unknown commercial terms, and unperformed live tests are distinguished.

2 reviewed sources · initial profile, capabilities, pricing mode, setup guidance, source review

OpenEvidence →

Profile added from fetched official sources. Vendor statements, proposed evaluation, unknown commercial terms, and unperformed live tests are distinguished.

1 reviewed sources · initial profile, capabilities, pricing mode, setup guidance, source review

Spellbook →

Profile added from fetched official sources. Vendor statements, proposed evaluation, unknown commercial terms, and unperformed live tests are distinguished.

2 reviewed sources · initial profile, capabilities, pricing mode, setup guidance, source review

Solve Intelligence →

Profile added from fetched official sources. Vendor statements, proposed evaluation, unknown commercial terms, and unperformed live tests are distinguished.

1 reviewed sources · initial profile, capabilities, pricing mode, setup guidance, source review

AGRIVI AI Engage →

Profile added from fetched official sources. Vendor statements, proposed evaluation, unknown commercial terms, and unperformed live tests are distinguished.

2 reviewed sources · initial profile, capabilities, pricing mode, setup guidance, source review

Hebbia →

Profile added from fetched official sources. Vendor statements, proposed evaluation, unknown commercial terms, and unperformed live tests are distinguished.

1 reviewed sources · initial profile, capabilities, pricing mode, setup guidance, source review

Dune →

Profile added from fetched official sources. Vendor statements, proposed evaluation, unknown commercial terms, and unperformed live tests are distinguished.

4 reviewed sources · initial profile, capabilities, pricing mode, setup guidance, source review

Cursor →

Profile added from fetched official sources. Vendor statements, proposed evaluation, unknown commercial terms, and unperformed live tests are distinguished.

2 reviewed sources · initial profile, capabilities, pricing mode, setup guidance, source review

GitHub Copilot →

Profile added from fetched official sources. Vendor statements, proposed evaluation, unknown commercial terms, and unperformed live tests are distinguished.

2 reviewed sources · initial profile, capabilities, pricing mode, setup guidance, source review

Replit Agent →

Profile added from fetched official sources. Vendor statements, proposed evaluation, unknown commercial terms, and unperformed live tests are distinguished.

2 reviewed sources · initial profile, capabilities, pricing mode, setup guidance, source review

Cline →

Profile added from fetched official sources. Vendor statements, proposed evaluation, unknown commercial terms, and unperformed live tests are distinguished.

2 reviewed sources · initial profile, capabilities, pricing mode, setup guidance, source review

OpenHands →

Profile added from fetched official sources. Vendor statements, proposed evaluation, unknown commercial terms, and unperformed live tests are distinguished.

1 reviewed sources · initial profile, capabilities, pricing mode, setup guidance, source review

Aider →

Profile added from fetched official sources. Vendor statements, proposed evaluation, unknown commercial terms, and unperformed live tests are distinguished.

2 reviewed sources · initial profile, capabilities, pricing mode, setup guidance, source review

Salesforce Agentforce →

Profile added from fetched official sources. Vendor statements, proposed evaluation, unknown commercial terms, and unperformed live tests are distinguished.

2 reviewed sources · initial profile, capabilities, pricing mode, setup guidance, source review

HubSpot Agent Hub →

Profile added from fetched official sources. Vendor statements, proposed evaluation, unknown commercial terms, and unperformed live tests are distinguished.

1 reviewed sources · initial profile, capabilities, pricing mode, setup guidance, source review

Lindy →

Profile added from fetched official sources. Vendor statements, proposed evaluation, unknown commercial terms, and unperformed live tests are distinguished.

2 reviewed sources · initial profile, capabilities, pricing mode, setup guidance, source review

Clay →

Profile added from fetched official sources. Vendor statements, proposed evaluation, unknown commercial terms, and unperformed live tests are distinguished.

2 reviewed sources · initial profile, capabilities, pricing mode, setup guidance, source review

Fin by Intercom →

Profile added from fetched official sources. Vendor statements, proposed evaluation, unknown commercial terms, and unperformed live tests are distinguished.

2 reviewed sources · initial profile, capabilities, pricing mode, setup guidance, source review

Zendesk AI Agents →

Profile added from fetched official sources. Vendor statements, proposed evaluation, unknown commercial terms, and unperformed live tests are distinguished.

2 reviewed sources · initial profile, capabilities, pricing mode, setup guidance, source review

Scite →

Profile added from fetched official sources. Vendor statements, proposed evaluation, unknown commercial terms, and unperformed live tests are distinguished.

2 reviewed sources · initial profile, capabilities, pricing mode, setup guidance, source review

SciSpace Research Agents →

Profile added from fetched official sources. Vendor statements, proposed evaluation, unknown commercial terms, and unperformed live tests are distinguished.

2 reviewed sources · initial profile, capabilities, pricing mode, setup guidance, source review

Khanmigo →

Profile added from fetched official sources. Vendor statements, proposed evaluation, unknown commercial terms, and unperformed live tests are distinguished.

2 reviewed sources · initial profile, capabilities, pricing mode, setup guidance, source review

NotebookLM →

Profile added from fetched official sources. Vendor statements, proposed evaluation, unknown commercial terms, and unperformed live tests are distinguished.

4 reviewed sources · initial profile, capabilities, pricing mode, setup guidance, source review

Microsoft Copilot Studio →

Profile added from fetched official sources. Vendor statements, proposed evaluation, unknown commercial terms, and unperformed live tests are distinguished.

2 reviewed sources · initial profile, capabilities, pricing mode, setup guidance, source review

Glean Agents →

Profile added from fetched official sources. Vendor statements, proposed evaluation, unknown commercial terms, and unperformed live tests are distinguished.

2 reviewed sources · initial profile, capabilities, pricing mode, setup guidance, source review

Dust →

Profile added from fetched official sources. Vendor statements, proposed evaluation, unknown commercial terms, and unperformed live tests are distinguished.

2 reviewed sources · initial profile, capabilities, pricing mode, setup guidance, source review

Dify →

Profile added from fetched official sources. Vendor statements, proposed evaluation, unknown commercial terms, and unperformed live tests are distinguished.

2 reviewed sources · initial profile, capabilities, pricing mode, setup guidance, source review

CrewAI →

Profile added from fetched official sources. Vendor statements, proposed evaluation, unknown commercial terms, and unperformed live tests are distinguished.

2 reviewed sources · initial profile, capabilities, pricing mode, setup guidance, source review

LangGraph →

Profile added from fetched official sources. Vendor statements, proposed evaluation, unknown commercial terms, and unperformed live tests are distinguished.

2 reviewed sources · initial profile, capabilities, pricing mode, setup guidance, source review

Pikpop →

User-requested profile added from the rendered official homepage. AI roles, task suitability, trial conditions and blocked preset demonstration are distinguished from independent performance testing.

1 reviewed sources · initial profile, capabilities, pricing offer, access attempt, source review

Postiz →

Added a focused founder-created or self-described small-team product from fetched official sources, distinguishing assistant, agent-connected workspace, automated workflow and specialist image tool. Live product output tests remain separate.

6 reviewed sources · initial profile, capabilities, independent or small-team background, pricing and access boundaries, setup guidance, source review

Typefully →

Added a focused founder-created or self-described small-team product from fetched official sources, distinguishing assistant, agent-connected workspace, automated workflow and specialist image tool. Live product output tests remain separate.

6 reviewed sources · initial profile, capabilities, independent or small-team background, pricing and access boundaries, setup guidance, source review

Taja AI →

Added a focused founder-created or self-described small-team product from fetched official sources, distinguishing assistant, agent-connected workspace, automated workflow and specialist image tool. Live product output tests remain separate.

3 reviewed sources · initial profile, capabilities, independent or small-team background, pricing and access boundaries, setup guidance, source review

Photo AI →

Added a focused founder-created or self-described small-team product from fetched official sources, distinguishing assistant, agent-connected workspace, automated workflow and specialist image tool. Live product output tests remain separate.

4 reviewed sources · initial profile, capabilities, independent or small-team background, pricing and access boundaries, setup guidance, source review

Goblin Tools →

Founder-built or specialist-vendor profile added from fetched official sources. Current headcount and controlling ownership remain unknown; planned product tests are not described as completed.

5 reviewed sources · initial profile, creator and operator provenance, capabilities, pricing and access boundaries, setup guidance, source review

Saner.AI →

Founder-built or specialist-vendor profile added from fetched official sources. Current headcount and controlling ownership remain unknown; planned product tests are not described as completed.

4 reviewed sources · initial profile, creator and operator provenance, capabilities, pricing and access boundaries, setup guidance, source review

Skyvern →

Founder-built or specialist-vendor profile added from fetched official sources. Current headcount and controlling ownership remain unknown; planned product tests are not described as completed.

6 reviewed sources · initial profile, creator and operator provenance, capabilities, pricing and access boundaries, setup guidance, source review

AudioPen →

Added a focused practical-tool profile with attributed sources, confirmed offer periods, edition and data-path limits, and two unexecuted product-specific cases. Current company size and ownership are not certified.

3 reviewed sources · initial profile, task fit, creator and operator provenance, capabilities, pricing periods and access boundaries, setup guidance, source review, product-specific evaluation plans

Chatbase →

Added a focused practical-tool profile with attributed sources, confirmed offer periods, edition and data-path limits, and two unexecuted product-specific cases. Current company size and ownership are not certified.

6 reviewed sources · initial profile, task fit, creator and operator provenance, capabilities, pricing periods and access boundaries, setup guidance, source review, product-specific evaluation plans

AnythingLLM →

Added a focused practical-tool profile with attributed sources, confirmed offer periods, edition and data-path limits, and two unexecuted product-specific cases. Current company size and ownership are not certified.

6 reviewed sources · initial profile, task fit, creator and operator provenance, capabilities, pricing periods and access boundaries, setup guidance, source review, product-specific evaluation plans

TypingMind →

Added a creator-origin practical client profile with nine retained read sources and two proposed cases. Amount symbols and periods are recorded without inventing an ISO currency; historical creator attribution does not establish current company size, ownership, autonomy or executed outcomes.

9 reviewed sources · initial profile, task fit, attributed creator and contracting operator, capabilities and context boundaries, quoted starting prices with unknown currency code, Personal versus Teams restrictions, setup guidance, source review, two unexecuted product-specific evaluation plans

Open WebUI →

Added Open WebUI as a practical creator-origin assistant with nine retained official sources, complete original primary/boundary scenarios and eight generated product-specific FAQs. Current headcount/control and mandatory software checkout price remain unknown. Absent tools in ordinary-chat Notes do not establish universal security, native RAG or executed outcomes.

9 reviewed sources · initial practical assistant profile, attributed creator, official small-core-team description and license holder, custom source-available branding license, preset versus model-weight training, ordinary-chat full-Note attachment versus note-editor tool path, null price amount, currency and billing unit, setup and Python release prerequisites, nine independently hash-verified source records, two original unexecuted product-specific evaluation plans

Voicenotes →

Added Voicenotes for daily recorded-note drafting/recall with unknown ISO pricing currency and unverified Basic Ask AI entitlement. Actual audio and native retrieval were not executed. Creator attribution does not establish present headcount/control; original case text and all conditions are preserved.

6 reviewed sources · initial practical AI profile, attributed creator/operator evidence and unknown current size/control, specific task inputs and outputs, current documented price and entitlement unknowns, step-by-step native setup and source/action boundaries, six retained hash-verified official sources, two original unexecuted research cases and eight generated FAQs

Pebblely →

Added Pebblely for small-shop product-image drafts with quoted US$ monthly plans, weaker co-founder attribution and no inferred current free credits. Actual upload/generation were not executed. Creator attribution does not establish present headcount/control; original case text and all conditions are preserved.

6 reviewed sources · initial practical AI profile, attributed creator/operator evidence and unknown current size/control, specific task inputs and outputs, current documented price and entitlement unknowns, step-by-step native setup and source/action boundaries, six retained hash-verified official sources, two original unexecuted research cases and eight generated FAQs

Skyvern →

Corrected the blanket PostgreSQL requirement using official v1.0.55 README and SDK/provider source. Embedded in-memory SQLite, persistent-server and cloud paths remain distinct. Installation readiness is not a product-case pass; original primary/boundary case definitions remain unchanged.

3 reviewed sources · local deployment paths, embedded SDK limits, Ollama vision configuration, setup and test access, fixed-version official sources

Rezi →

Added Rezi for personal resume drafting and target-role gap review. Official Jacob Jacquet/May 2015 attribution is separate from unknown current headcount/control; dollar currency and actual guest/AI/export entitlement remain unverified. Both synthetic fact-fidelity and missing-qualification cases are unexecuted. No candidate data, application or native generation was submitted.

4 reviewed sources · initial personal resume profile, HR and recruiting task discovery, attributed founder and unknown current size/control, documented plan prices with unknown dollar currency, guest documentation versus static signup redirects, four retained original official response bodies, two strict unexecuted resume cases and eight product FAQs, official original PNG identity asset

Open Interpreter →

Added the current Rust Open Interpreter for exact local CSV/file tasks. Legacy Killian Lucas attribution does not certify present team size or ownership. One native startup reached zero model calls and no output, then further execution stopped because shared Windows skills discovery could not be isolated through supported configuration. Both complete original cases remain unexecuted; existing registered product results are unchanged.

11 reviewed sources · creator-origin local file/data agent profile, current Rust versus legacy Python distinction, eleven retained official source responses, two frozen native cases and eight FAQs, official SVG identity, pre-model startup block and public metadata evidence

Rembg →

Added Rembg for local product-image cutouts with official Daniel Gatis creator attribution; current team size and control remain unknown. Two original native HTTP cases passed eight pre-existing conditions with checksum-verified u2netp and explicit putalpha=true. The unedited outputs, original fixtures and masks, multipart requests, metrics and native provenance remain downloadable. Default BRIA, Gradio interaction, cloud inference and broad natural-photo quality are outside the registered scope. No official favicon was verified; the existing generic Bot fallback is retained.

11 reviewed sources · creator-origin product-image background-removal profile, specialist AI tool classification and unknown current ownership, eleven retained official source responses, two complete frozen native HTTP image cases and eight FAQs, explicit u2netp/putalpha configuration and model/software license distinction, generic existing brand-icon fallback without invented vendor logo, registered native image inference and before/after PNG evidence

Khoj →

Added the current self-hosted Khoj profile with an explicit April 15, 2026 Cloud sunset. Frozen document-answer cases are unexecuted; preparation does not establish answer quality. Current headcount and corporate control remain unknown.

8 reviewed sources · initial profile, current cloud sunset, self-hosted route, creator sources, two frozen cases, eight FAQ

Vibe →

Added creator-origin Vibe with official release, local transcription and optional online analysis clearly separated. Two original transcription cases were executed once each and each failed one unchanged strict output condition; current headcount and control remain unknown.

11 reviewed sources · initial profile, creator attribution, local transcription versus optional analysis, two frozen cases, eight FAQ

Khoj →

Registered two completed original native self-hosted answers with the fixed tiny local Qwen model: primary passed one of four conditions and boundary passed two of four; both cases failed. Original questions, three synthetic documents, references and all eight criteria remain unchanged. Retained an earlier controller interruption separately from completed quality results, disclosed one unchanged primary environment recovery and cumulative usage uncertainty, and preserved the observed non-loopback connection limitation. Correct retrieved passages did not establish correct answer grounding. No cloud, stronger-model or overall product-quality result is claimed.

4 reviewed sources · native execution evidence, actual case results, registry-derived test FAQ, incomplete attempt history, runtime and network limitations

Vibe →

Clarified that both native cases completed with the exact fixed model while the selected CPU/GPU backend remains unmeasured. Workflow and dependency lists now identify retained pre-resource planning text beside that text; original case objects, conditions, outputs and evidence bytes are unchanged.

2 reviewed sources · fixed-model inference compatibility, nearby historical planning labels

LLM →

Added Simon Willison’s LLM CLI as a Specialist AI tool for authorized text-to-JSON drafts. Retained eleven official source-access rows, creator attribution with current team/ownership unknown, eight detailed FAQs and two unchanged original cases. Both real cases completed once through the full official 0.36 console using Ollama 0.35.0 and fixed local Qwen weights: primary 7/8, boundary 4/8, 11/16 conditions passed and both cases failed. Preserved the official system trailing-LF mapping failure, untrusted fulfillment values and unsupported order/email claims, the complete failed C07 despite its identifier-only subcheck passing, seven preparation CLI calls and three runtime starts. Public evidence contains the unchanged original input/output bytes and complete safe rule reviews. No quality retry, output repair, general accuracy score, commercial-rights or zero-cost claim.

11 reviewed sources · independent creator profile, native original case results, public evidence, practical collection, search metadata, scope and cost limitations

LibreTranslate →

Added LibreTranslate as a Specialist AI translation tool for authorized notes and product drafts, with twelve retained official sources and unknown present team/operator/ownership. Complete official local GUI/backend Spanish-to-English cases each sent one POST and passed all four unchanged original conditions, 8/8 total; only first native translatedText was scored and all three alternatives were retained. Original inputs/responses/visible outputs/screenshots and safe independent review are public. Copy UI reported copied but the session clipboard did not match; clipboard content was not retained and success remains unverified, separately from the eight criteria. Startup timeouts, dependency/queue warnings, runtime/software/model identities and distinct OPUS/Argos/Stanza/MiniSBD attribution remain disclosed. No broader translation-quality, business-action, commercial-rights or zero-cost certification.

12 reviewed sources · creator project profile, native local GUI case results, safe original evidence, model attribution, copy limitation, practical collection, search metadata

Photopea →

Added Photopea as a Specialist AI tool for individuals, creators and small sellers, with nine verified official source bodies, dated Ivan Kutskir creator attribution and unknown current team/operator/ownership. Explained free editor terms, optional Premium, transparent PNG workflow and browser embedding API, while keeping all AI processing, model/build, price/credit and live-command gates unverified. Preserved both v2 original frozen cases and all twelve pass conditions verbatim. Both cases remain not_executed with authorization pending: no native image import, AI request, export, score or general image-quality claim is recorded.

9 reviewed sources · individual-developer origin, practical browser editor profile, official source review, eight detailed FAQs, unchanged frozen test cases, pending authorization and unexecuted status

Kokoro Web →

Added Kokoro Web as an individual-author Specialist AI tool for script-to-audio narration. Retained source bytes support Eduardo Lat / Luis Eduardo attribution, MIT app and Apache-2.0 model licenses, dated maintenance and Browser/native MP3 workflow. Both original case inputs and all twelve conditions remain verbatim. Native execution facts follow the root handoff; listening conditions remain unverified and no score or broader speech-quality claim is made. Both original native results are partial because listening checks remain unverified.

11 reviewed sources · individual author and Person attribution, pure text-to-speech profile, maintenance snapshot, retained official-source evidence, eight detailed FAQs, unchanged original two-case conditions, partial native MP3 execution and unverified listening conditions

Scribe OCR →

Added Scribe OCR as an individual-author Specialist AI tool for document image-to-text and searchable PDF tasks. Nineteen retained official sources support Balearica Person attribution, AGPL-3.0 application licensing, complete browser import/recognition/download workflows, and the distinct optional AWS Textract cloud mode. Both original synthetic label inputs and all twelve original conditions remain unchanged. Prepared, not executed: specific two-PNG import permission remains pending, actual runtime contract is unfrozen, and no import, OCR, output, export, result score or execution record is created. Real name, current team size, legal operator, hosted-build identity and independent privacy verification remain unknown.

19 reviewed sources · individual creator and Person attribution, browser OCR and proofreading profile, retained official source evidence, eight detailed FAQs, unchanged two-case input and acceptance contract, prepared not executed status, practical collection and search metadata

Chatbox Community Edition →

Added Chatbox Community Edition as a practical Specialist AI tool and AI conversation client with retained Benn Huang individual-author evidence. Seventeen official sources support the fixed GPLv3 CE source, complete standalone web renderer, provider guides, browser storage and native exporter. Both original small-shop inputs, twelve pass conditions and four failure conditions are unchanged. Two native text cases completed with one actual backend generation and the retained first native Markdown export for each case; both failed in the recorded Qwen2.5-Coder-1.5B-Instruct Q4_K_M configuration. The pickup draft omitted the non-stacking rule, missed the required two-decimal coupon price and reversed checklist dates; the ceramic reply made unsupported food-contact, dishwasher and delivery guarantees. Original output quality was not retried. Web RAG, MCP, agent mode, skills, Pro Work Mode and external actions were excluded. Original condition judgements were pickup 3 passed/3 failed and ceramic 2 passed/4 failed. Actual requests retain native Additional info / Runtime system context with model alias, session start time and system-reminder guidance despite the visible Inject metadata setting being off. Runtime source commit is fixed; the About version number was blank, while official release v1.23.5 and root package 0.0.1 are separate metadata. Current team size, legal operator and controlling ownership remain unknown.

17 reviewed sources · individual author and Person attribution, AI conversation-client profile, local provider and native export workflow, retained official-source evidence, eight detailed FAQs, unchanged two-case original inputs and twelve conditions, two native completed failed cases and original first Markdown exports, small-shop task, practical collection and authored search metadata

SiYuan →

Added SiYuan as a practical Specialist AI tool for supervised note summaries and publication checklists. Fixed 3.8.6 source and official privacy text identify Yunnan Liandi Technology Co., Ltd.; the package names Liang Ding and Liyuan Li as maintainers, without establishing current headcount or controlling ownership. Nineteen retained official sources support local storage, AGPLv3, provider configuration, inline editing and distribution-specific limitations. Both original synthetic notes, exact action templates, twelve acceptance conditions and four failure conditions remain unchanged. Cases are prepared but unexecuted; no completed model response, score, browser insertion or file export is claimed. Agent Chat, MCP, web search, cloud sync, embedding, image generation and external actions are excluded from the defined endpoint pilot.

19 reviewed sources · company developer and named maintainers, local knowledge-workspace and supervised AI-editing profile, nineteen retained official sources, eight detailed FAQs, unchanged two original note inputs and editing actions, twelve original acceptance conditions, personal-note synthesis task and practical collection, Organization and authored search metadata

SiYuan →

Registered both original SiYuan 3.8.6 native editor cases as executed and failed with Qwen2.5-Coder-1.5B-Instruct Q4_K_M. Eight conditions passed, three failed and one remained unverified. Primary preserved its main facts but missed the requested heading structure. Boundary retained confirmed facts but copied contaminated candidates into the review section and asserted invitations were already sent; no actual sending is evidenced. Two real model forwards, zero quality reruns and unchanged native note exports were retained. The wrong-route HTTP404 occurred before model processing; original payload bytes were recovered at the official route without rearming. Browser insertion, generated-file export, Agent Chat, MCP, skills and cloud quality are untested.

2 reviewed sources · first native editor responses, unchanged original condition verdicts, native request and stream provenance, readonly synthetic notes, observed runtime limitations

BgSub →

Added BgSub as a browser cutout tool for independent sellers and creators. An official 2022-08-21 article documents its independently developed origin and the public author alias DeepKnight; real name, current legal operator, team size and controlling ownership remain unknown. Fourteen retained official responses support documented capabilities and route boundaries, with entry shells marked limited. Two original 512 x 512 synthetic cases remain prepared and unexecuted: no import, automatic result, native PNG download or pixel score is claimed. All seven fields, twelve acceptance conditions and six failure conditions are preserved. Free processing and maximum size are vendor statements; current account/full-size entitlement is unverified. The project uses neutral Thing structured data and the existing generic icon.

14 reviewed sources · independent-development background-removal profile and Specialist AI tool classification, fourteen retained official source responses with app-shell limits, native cutout workflow, current entitlement unknowns and separate privacy paths, two original frozen seven-field cases with twelve acceptance and six failure conditions, eight individually written FAQs and authored search metadata, neutral project Thing and existing generic brand fallback

SillyTavern →

Prepared a SillyTavern profile for original character dialogue and story-bible review. Official docs describe a community passion project and independent development; named maintainers do not establish a legal company, headcount or controlling ownership. The stable 1.19.0 tag is pinned to 7e8663cd9c184a550b37238218bdd32c6efc68e9. Both original Mara cases remain unexecuted in this proposal, with seven fields per case, twelve acceptance conditions and six failure conditions preserved exactly. World Info, embeddings, extensions and external tools are excluded from the two-case pilot. No model response, score, export success or business outcome is claimed.

17 reviewed sources · community-origin character-writing profile and Specialist AI tool classification, seventeen retained official source responses and exact stable tag identity, native character workflow, model/provider costs and AGPL boundaries, two original frozen seven-field cases with twelve acceptance and six failure conditions, eight individually written FAQs with standard coverage question, authored search metadata and existing generic brand fallback

SillyTavern →

Registered both original SillyTavern 1.19.0 character-chat cases as executed and failed with Qwen2.5-Coder-1.5B-Instruct Q4_K_M. Eight conditions passed, four failed and none remained unverified. Primary kept the five canon facts and lantern uncertainty, but its 143 spoken words and Mara-plus-Ivo script failed the requested voice/word-range condition, and Writer review was absent. Boundary kept exactly two unknown-detail bullets and made no action claim, but omitted the prop-conflict explanation and all required canon facts. Each case used a fresh native conversation, one actual model forward and its unchanged first response; there were zero quality reruns. Two CONNECT attempts to aihorde.net:443 were rejected by the task application proxy, with zero proxy forwards; this does not prove OS isolation or all-app privacy. World Info, embedding, extensions, tools, cloud models, native export downloads and publication remain untested. Results cover only the recorded local configuration.

3 reviewed sources · first native character-chat responses and unchanged server-created JSONL, twelve original condition verdicts, native provider request and response provenance, observed local-model and application-proxy limitations

big-AGI →

Published the big-AGI profile with both original native Beam-and-standard-Fuse cases completed once and failed using the complete main snapshot 80d366a88cc8 (package 2.1.1) and cached Qwen2.5-Coder-1.5B-Instruct Q4_K_M. Primary passed three of six conditions; boundary passed one of six; four passed, eight failed and none remained unverified. Five of six failure conditions were triggered. Primary missed Script, the one-shirt-per-attendee rule and exact Owner review heading. Boundary omitted the correction body and workshop name, adopted conflicting capacity/item/guarantee fields, used eighteen bullets rather than two and asserted booking/email completion without real external actions. Each fresh conversation used two first same-model rays plus one first standard Fuse, six actual forwards total and zero quality reruns. Native max_tokens was omitted, Context Override was 8192 and token counting was approximate. Read-only native IndexedDB recovery retained exact Persona, user and accepted first Fuse; the UI Download JSON action did not complete. Captured AIX cancellations remain documented alongside complete provider stop/DONE streams. Same-model alternatives do not establish cross-model accuracy or overall product quality.

4 reviewed sources · complete native Beam and first standard Fuse responses, twelve frozen condition verdicts and six failure conditions, read-only native IndexedDB history recovery, recorded model and original runtime limits

Subtitle Edit →

Published the Subtitle Edit profile with both original official SeConv v5.2.0 Windows x64 native SRT cases completed once and failed using the cached Qwen2.5-Coder-1.5B-Instruct Q4_K_M model. Primary passed four of six conditions and failed PC3/PC4: the first native file retained four original cues and exact timecodes, the shop name, Arabic numerals, 12 dollars and the no-order prohibition, but translated 2 pens as 2 pencils and changed pickup to arrival without retaining collection meaning. Boundary passed three of six conditions and failed PC3/PC4/PC5: literal [ORDER_ID] and みどり文具 were replaced, Arabic 2 became the Chinese numeral 两, and placeholder remained English. Both pickup-date/profit unknowns, the review-only limitation and prohibitions on email sending/refund promises remained; no completed business action was claimed. Seven of twelve original conditions passed, five failed and none remained unverified; two of six failure conditions were triggered. The tally includes native workflow, structure and evidence conditions and is not a translation-accuracy percentage. The selected complete official SeConv v5.2.0 Windows x64 subcomponent used the cached Qwen2.5-Coder-1.5B-Instruct Q4_K_M backend through native llamacpp transport. There was one native conversion and one actual provider forward per case, two forwards total, with no observed internal fallback, quality rerun or manual output repair. Native file creation and original cue/timecode preservation passed; unsupported business completion was not claimed. The actual subtitle files and original complete case contract were retained unchanged. These results cover only this original local configuration. Desktop GUI, ASR, dubbing, OCR, subtitle burn-in, final-video synchronization, upload and publication were not tested. Translating the boundary wording does not measure autonomous agent refusal or action permissions.

3 reviewed sources · source-backed profile and individual creator attribution, complete official SeConv first native subtitle files, twelve frozen condition verdicts and six failure conditions, registered cached model and original runtime limits

Fabric →

Added Fabric with detailed native content-pattern setup, costs, limits and seven retained official sources. Both original create_micro_summary cases ran once through official v1.4.507 with the cached local Qwen model and failed: primary 4/6 and boundary 2/6. Native exports completed; requested formatting failed, and the corrected newsletter mixed superseded targets and an unsupported profit expectation. Full first outputs and all original conditions remain available; the tally is not summary accuracy.

3 reviewed sources · Source-backed content pipeline and Daniel Miessler creator attribution, Two first official native micro-summary outputs and complete original condition verdicts, Person creator schema, task links and public evidence

Open Interpreter →

Registered both original Rust0.0.55 Windows native CSV cases after one real local-model response per case. Both failed with the cached Qwen2.5-Coder1.5B model: no tools ran or summary.csv was created; primary totals were wrong and the boundary answer falsely claimed file creation. Of six original conditions, two passed, three failed and one remained unverified. Native fallback model metadata and full-disk-read limits are disclosed. The actual note was not read, so injection resistance is unverified. Official source dates and all24 earlier runs are retained. The existing practical collection now states its small-business and solo-work uses; no new duplicate keyword pages or overall performance ranking was added.

3 reviewed sources · Two completed first original native CSV cases and six unchanged pass-condition verdicts, Missing generated file, wrong primary totals and fabricated boundary file-completion claim, Cached model and native fallback-metadata provenance with downloadable first outputs, Source-preserving profile update and practical small-business collection wording

AIChat →

Added individually attributed AIChat with fixed native binary, dual licenses, two original once-run drafting cases and eight FAQs. Both full contracts failed; ordinary has a format-only failure, boundary also adds a wrong UTC offset and misses memo rejection. Actual temperature omission is disclosed. No paid provider, weights or business connection was added.

7 reviewed sources · initial profile, creator attribution, official sources, native first-output cases, FAQ, search metadata, practical collection

ShellGPT →

Added personally attributed ShellGPT with 9 retained official sources, 8 FAQs and two once-run native command-generation cases. Both complete contracts failed; fences, files-only filter, filename projection and post-output console error remain visible. Only static inspection, no command execution or quality rerun.

9 reviewed sources · initial profile, creator attribution and disclosed affiliation, official sources, native original cases, FAQ, search metadata, practical collection

tgpt →

Added tgpt with 8 retained official sources, 8 FAQ entries and two once-run native stdin cases. Both cases fail: missing next-update commitment in primary; fabricated memo tracking number and fences in boundary. No quality rerun or business action.

8 reviewed sources · initial profile, official sources, creator and maintenance attribution, native original cases, FAQ, search metadata, practical collection

Mods →

Added Mods with 9 retained official sources, 8 FAQ entries and two once-run native stdin cases. Both cases fail stock classification despite valid native JSON; boundary adopts memo-selected E-50. No quality rerun or business action.

9 reviewed sources · initial profile, official sources, creator and maintenance attribution, native original cases, FAQ, search metadata, practical collection

Chatblade →

Added Chatblade with 9 retained official sources, 8 FAQ entries and two once-run native stdin cases. Primary passes; boundary fails null semantics while preserving confirmed values and avoiding memo changes. No quality rerun or business action.

9 reviewed sources · initial profile, official sources, creator and maintenance attribution, native original cases, FAQ, search metadata, practical collection

Aider →

Both original whole-file editing cases pass all frozen conditions. Legacy Aider cases remain unchanged; gptme has no model output.

10 reviewed sources · official sources, native task records, runtime limits, FAQ, creator attribution

gptme →

Both native task attempts fail before generation; model response quality is unverified. Legacy Aider cases remain unchanged; gptme has no model output.

10 reviewed sources · official sources, native task records, runtime limits, FAQ, creator attribution

NotebookLM →

Primary fails the frozen conflict-table expectation despite correct sourced facts and a conflict list; boundary passes missing-policy handling. Original case definitions retained; observed native UI and backend limits disclosed.

1 reviewed sources · native original case evidence, expected-output review, FAQ, observed native UI, test access, search metadata

GPTScript →

Both original no-tools inquiry contracts fail: fenced JSON in both; boundary additionally adopts untrusted memo facts. Archived repository and Acorn Labs source credit disclosed.

9 reviewed sources · initial profile, official sources, native original cases, first task results, runtime limits, FAQ, source attribution

OpenCode →

Both original contracts fail with no inventory JSON: frozen steps=1 injects a native maximum-steps prompt on the first request, returned as the first text. Default coding-agent quality remains untested.

11 reviewed sources · initial profile, official sources, original native cases, configuration-confounded first results, FAQ, source attribution

Goose →

Both no-extension return JSON contracts fail: fenced output in both; boundary also adopts memo units 9 and deadline tomorrow.

10 reviewed sources · initial profile, official sources, original native cases, first results, FAQ, AAIF source affiliation

Crush →

Both original invoice JSON contracts fail: fenced task output in both; boundary also adopts memo total 99 and currency USD. Two task and two auxiliary title calls are recorded separately.

10 reviewed sources · initial profile, official sources, original native cases, first task/title evidence, FAQ, FSL license scope, shared Charm profile retention

Pi →

Both native event-row contracts fail from fenced assistant text; boundary also adopts remaining_seats 0 and host Aster. First assistant is distinct from the native JSON event wrapper.

9 reviewed sources · initial profile, official sources, original native cases, first results and event projections, FAQ, personal attribution/current repository scope

nanobot →

Native Python AgentLoop primary closure notice passes; boundary fails by adopting an unsupported reopening date. Caller-owned registry removal is distinct from CLI permissions.

11 reviewed sources · initial profile, official sources, original native SDK cases, first outcomes and SDK caller evidence, FAQ, SDK versus CLI scope

PicoClaw →

Both native CLI shop-notice originals pass, including the copied reopening memo. Same-task comparison with nanobot favors PicoClaw in this two-scenario recorded pilot.

11 reviewed sources · initial profile, official sources, first native CLI cases, same-task comparison, FAQ

Muse →

Official personal assistant profile added. Neither native task case executed; no performance score or professional review claimed.

2 reviewed sources · initial profile, official sources, brand icon and official cover, personal admin task, unexecuted evaluation plans, FAQ

Grok Bot →

Official personal assistant profile added. Neither native task case executed; no performance score or professional review claimed.

3 reviewed sources · initial profile, official sources, brand icon and official cover, personal admin task, unexecuted evaluation plans, FAQ

Dots →

Added sourced OpenAI Dots personal-agent profile. No native product execution, rating or independent professional/security review claimed.

3 reviewed sources · initial profile, official character and cover artwork, personal and creator tasks, unexecuted plans, eight FAQs

Goblin Tools →

Executed both original Magic ToDo cases once in the official public UI. Both failed fixed conditions: inconsistent and over-budget estimates in the primary; a work-budget cutoff treated as finish time without qualifying completion in the boundary. First text and screenshots retained. Source-review dates remain unchanged; no general accuracy score or human review.

1 reviewed sources · native product tests, first outputs, failure evidence

Scribe OCR →

Executed two original PNG labels once each after settings were frozen. Clear label passed 6/6 conditions; low-contrast rotated label failed exact SKU, price and searchable-PDF conditions (3/6 passed). First TXT character error: 0/219 and 3/244. Original image pixels retained; boundary PDF syntax warnings preserved. Source-review dates remain unchanged; this is a two-sample single-agent test, not general accuracy.

1 reviewed sources · native OCR tests, first TXT and PDF exports, failure evidence

Photopea →

Registered first native public-UI results: the synthetic bottle passed 6/6 conditions and the low-contrast mug failed, with 3/6 passed. Retained the original frozen contract and every original case field, pre-inference URL-import addendum, inputs/masks, first exports and exact metrics. Backend model/provider, local/private inference and total cost remain unobserved. Historical official-source check dates are unchanged.

1 reviewed sources · two original native cases executed, first native PNG evidence, explicit pre-inference URL-import deviation, passed bottle and failed low-contrast mug, current test-access, search description and FAQ status

BgSub →

Executed both original 512x512 synthetic cases once after explicit Terms approval. Both failed, each passing 3/6 fixed conditions: the bottle lost most of its cream label and the mug missed body-opacity and handle-hole-transparency thresholds despite passing overall silhouette IoU. Retained original cases, first automatic results, native PNGs, evaluator-only masks and exact failed conditions. No login, purchase, FullSize, contribution upload, manual repair or quality rerun; model, inference runtime and total cost remain unknown. Historical official-source check dates are unchanged.

1 reviewed sources · explicit Terms approval and two original native cases executed, first native automatic PNG exports, both synthetic cases failed with 3/6 conditions passed each, current test-access, search description and FAQ status

PackyAPI →

Site-owner-requested model API relay profile added with official identity artwork, model-dependent pricing fields, terms and a separate catalogue-only check. API inference remains untested. The supplied referral registration link is disclosed and sponsored. The initial documentation connection error is retained as history; the overview and quick-start guide were read after access recovered.

3 reviewed sources · initial profile, official sources, catalogue UI check, referral disclosure, same-day documentation recovery

Brandfolia →

Recorded one fixed public-sample primary case: 3/3 conditions passed, with 30 PDF pages and 17 SVGs inspected from the first source ZIP. Separate PDF-button download unconfirmed. Boundary, paid generation and checkout untested; same-owner internal evaluation disclosed.

2 reviewed sources · sample edit/export evaluation, first native package evidence, same-owner scope limits

Brandfolia →

Added a disclosed same-owner specialist branding tool from official sources. Paid generation, checkout and export quality remain untested; no referral link or independent endorsement.

3 reviewed sources · initial profile, official sources, same-owner disclosure