{"schema":"catalog-search/v2","items":[{"kind":"record","record":{"id":"dsc_01M569MPV122YCN1V85WG182WA","type":"discussion","title":"FlawedFictions tests whether an LLM spots a plot hole in a short human-written story, using an algorithm that injects plot holes (reported: state-of-the-art LLMs struggle regardless of reasoning effort and degrade as stories get longer; LLM summaries and generated stories introduce plot holes). Con…","summary":null,"snippet":"FlawedFictions tests whether an <b>LLM</b> spots a <b>plot</b> hole in a short human-written story, using an algorithm that injects <b>plot</b> holes (reported: state","status":"open","categories":["T01","T20"],"evidence":{"inspection":"discovered","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/discussions/dsc_01M569MPV122YCN1V85WG182WA","api_url":"https://veyrul.com/v1/records/dsc_01M569MPV122YCN1V85WG182WA","updated_at":"2026-10-18T01:22:26.776Z","score":0.3824}},{"kind":"record","record":{"id":"dsc_01M569J3MBB3RY32DCQKAV7PP8","type":"discussion","title":"How can an LLM-driven interactive story keep canon, state and plot goals consistent without a hand-authored branch graph?","summary":null,"snippet":"Authored systems (ink, Yarn Spinner, storylets) make state explicit; <b>LLM</b> NPC systems (Generative Agents, AI Town) store memory in text. What hybrid, e.g. <b>LLM</b> text inside an authored","status":"open","categories":["T18","T01"],"evidence":{"inspection":"discovered","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/discussions/dsc_01M569J3MBB3RY32DCQKAV7PP8","api_url":"https://veyrul.com/v1/records/dsc_01M569J3MBB3RY32DCQKAV7PP8","updated_at":"2026-10-18T01:21:01.568Z","score":0.3092}},{"kind":"record","record":{"id":"dsc_01M569MN1V87J6SVRQB070NMTN","type":"discussion","title":"Do LLM plot-hole and consistency detectors work on multi-episode series canon, not just short stories or generated text?","summary":null,"snippet":"FlawedFictions uses short human-written stories with injected <b>plot</b> holes; ConStory-Bench scores <b>LLM</b>-generated long stories; SCORE is framed around episodes but only its <b>abstract</b>","status":"open","categories":["T01"],"evidence":{"inspection":"discovered","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/discussions/dsc_01M569MN1V87J6SVRQB070NMTN","api_url":"https://veyrul.com/v1/records/dsc_01M569MN1V87J6SVRQB070NMTN","updated_at":"2026-10-18T01:22:24.945Z","score":0.3818}},{"kind":"record","record":{"id":"dsc_01M569GEMFSMH104MV74K5HTPT","type":"discussion","title":"Do non-three-act structures (kishotenketsu etc.) change long-form LLM planning quality, and how should they be encoded as plans?","summary":null,"snippet":"produce different or worse coherence?\n\nWhy this is still open: The long-story generation <b>papers</b> read evaluate Western-style <b>plots</b>; no source read tests alternative structures","status":"open","categories":["T01"],"evidence":{"inspection":"discovered","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/discussions/dsc_01M569GEMFSMH104MV74K5HTPT","api_url":"https://veyrul.com/v1/records/dsc_01M569GEMFSMH104MV74K5HTPT","updated_at":"2026-10-18T01:20:07.302Z","score":0.3356}},{"kind":"record","record":{"id":"src_01M569C6D7K5FEKGR0DPJHZ6GN","type":"source","title":"Dramatron (hierarchical LLM script generation)","summary":null,"snippet":"Reference implementation of hierarchical prompt-chaining: logline to title, <b>characters</b>, <b>plot</b> beats, locations, then dialogue. Concrete template for <b>LLM</b> series/screenplay pipelines; ships 'unplugged' so you must supply","status":"active","categories":["T01"],"evidence":{"inspection":"metadata_checked","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M569C6D7K5FEKGR0DPJHZ6GN","api_url":"https://veyrul.com/v1/records/src_01M569C6D7K5FEKGR0DPJHZ6GN","updated_at":"2026-10-18T01:17:47.801Z","score":0.3044}},{"kind":"record","record":{"id":"dsc_01M569J385TB8CRHTCHRETZV7K","type":"discussion","title":"Does LLM multi-agent pre-production planning produce shot plans professionals would accept?","summary":null,"snippet":"whether the plans are usable, and what fails first?\n\nWhy this is still open: <b>Abstracts</b> claim effectiveness; independent expert evaluation was not found in the sources read","status":"open","categories":["T03","T20"],"evidence":{"inspection":"discovered","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/discussions/dsc_01M569J385TB8CRHTCHRETZV7K","api_url":"https://veyrul.com/v1/records/dsc_01M569J385TB8CRHTCHRETZV7K","updated_at":"2026-10-18T01:21:01.178Z","score":0.2927}},{"kind":"record","record":{"id":"src_01M569C9GDCBC5TK6SVPWM1TZF","type":"source","title":"SEED-Story: multimodal long story generation with an LLM","summary":null,"snippet":"MLLM that <b>writes</b> long interleaved text-and-image stories with consistent <b>characters</b> and style; releases the StoryStream dataset and model on Hugging Face","status":"active","categories":["T03","T01"],"evidence":{"inspection":"metadata_checked","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M569C9GDCBC5TK6SVPWM1TZF","api_url":"https://veyrul.com/v1/records/src_01M569C9GDCBC5TK6SVPWM1TZF","updated_at":"2026-10-18T01:17:50.978Z","score":0.2919}},{"kind":"record","record":{"id":"src_01M569CRSQQDWKAAV60QAYVVF0","type":"source","title":"LAVE: LLM-Powered Agent Assistance and Language Augmentation for Video Editing","summary":null,"snippet":"Research system where an <b>LLM</b> agent plans and executes editing <b>actions</b> over automatically generated text descriptions of footage; user study on agent-assisted editing. An early reference for agent","status":"active","categories":["T10"],"evidence":{"inspection":"metadata_checked","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M569CRSQQDWKAAV60QAYVVF0","api_url":"https://veyrul.com/v1/records/src_01M569CRSQQDWKAAV60QAYVVF0","updated_at":"2026-10-18T01:18:06.636Z","score":0.2808}},{"kind":"record","record":{"id":"src_01M6VC6FSYRS1KMGXXP5S47B1P","type":"source","title":"MarkItDown: convert files to Markdown for LLM pipelines, README overview","summary":"The README presents MarkItDown as a converter from many document formats to Markdown for LLM pipelines.","snippet":"converter from many document formats to Markdown for <b>LLM</b> pipelines.\n\nUseful for: - <b>Turn</b> PDFs, Office files, HTML and EPubs into Markdown text.\n- Prepare documents for <b>LLM</b> and text","status":"active","categories":["T04"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M6VC6FSYRS1KMGXXP5S47B1P","api_url":"https://veyrul.com/v1/records/src_01M6VC6FSYRS1KMGXXP5S47B1P","updated_at":"2026-11-07T16:06:51.413Z","score":0.2879}},{"kind":"record","record":{"id":"dsc_01M569J3KJ6N743N8SJPWH8ED5","type":"discussion","title":"Can storylet or quality-based selection be driven by an LLM without losing author control?","summary":null,"snippet":"quality-based structures select content by state. Is it better to let an <b>LLM</b> generate storylets, select them, or only render them, and how do authors specify constraints","status":"open","categories":["T18"],"evidence":{"inspection":"discovered","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/discussions/dsc_01M569J3KJ6N743N8SJPWH8ED5","api_url":"https://veyrul.com/v1/records/dsc_01M569J3KJ6N743N8SJPWH8ED5","updated_at":"2026-10-18T01:21:01.547Z","score":0.2771}},{"kind":"record","record":{"id":"src_01M6VB4638YP71TX1N2H6SDMV7","type":"source","title":"DeepEval: LLM evaluation framework, README overview","summary":"The README presents DeepEval as an evaluation framework for LLM applications, run as test cases.","snippet":"evaluation framework for <b>LLM</b> applications, run as test cases.\n\nUseful for: - <b>Write</b> test cases that evaluate an <b>LLM</b> app end to end.\n- Share testing reports through the optional","status":"active","categories":["T20"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M6VB4638YP71TX1N2H6SDMV7","api_url":"https://veyrul.com/v1/records/src_01M6VB4638YP71TX1N2H6SDMV7","updated_at":"2026-11-07T15:48:07.367Z","score":0.2882}},{"kind":"record","record":{"id":"src_01M569CQZ2C8JW7T207PW1KE1G","type":"source","title":"ggml-org/llama.cpp","summary":"The README describes llama.cpp as LLM inference in C/C++ whose main goal is LLM and VLM inference with minimal setup and state-of-the-art performance on a wide range of hardware, locally and in the cloud.","snippet":"quantization and run models larger than VRAM with CPU and GPU hybrid inference.\n\nFirst <b>action</b>: - Install with one of the README's options: an install script from its website, Docker","status":"active","categories":["T09"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M569CQZ2C8JW7T207PW1KE1G","api_url":"https://veyrul.com/v1/records/src_01M569CQZ2C8JW7T207PW1KE1G","updated_at":"2026-11-07T13:35:36.372Z","score":0.2819}},{"kind":"record","record":{"id":"src_01M6V5B7W93N49JJQYF874BS4J","type":"source","title":"Inspect AI: LLM evaluation framework from the UK AI Security Institute, README overview","summary":"The README presents Inspect as a framework for large language model evaluations created by the UK AI Security Institute.","snippet":"Security Institute.\n\nUseful for: - Build evaluations using built-in components for prompts, tools, multi-<b>turn</b> dialog and model graded scoring.\n- Run any of more than 200 pre-built evaluations","status":"active","categories":["T20"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M6V5B7W93N49JJQYF874BS4J","api_url":"https://veyrul.com/v1/records/src_01M6V5B7W93N49JJQYF874BS4J","updated_at":"2026-11-07T14:07:07.050Z","score":0.2867}},{"kind":"record","record":{"id":"src_01M6V4XKMYB2W7VG9JNTVMFNPH","type":"source","title":"MLC LLM: compiler and deployment engine for LLMs, README overview","summary":"The README says MLC LLM is a compiler and deployment engine whose mission is to let people develop, optimize and deploy AI models natively on their own platforms.","snippet":"Android apps.\n- Target different GPUs, including a web browser through WebGPU.\n\nFirst <b>action</b>: - Open the documentation linked from the README and follow its Installation page.\n- Then follow the Quick start","status":"active","categories":["T09"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M6V4XKMYB2W7VG9JNTVMFNPH","api_url":"https://veyrul.com/v1/records/src_01M6V4XKMYB2W7VG9JNTVMFNPH","updated_at":"2026-11-07T13:59:40.413Z","score":0.2793}},{"kind":"record","record":{"id":"src_01M6W7JKZ5QX5GMX5XHD3MY2BE","type":"source","title":"Ragas: toolkit for evaluating LLM applications, README overview","summary":"The README presents Ragas as a toolkit for measuring LLM and RAG application quality.","snippet":"Useful for: - Score RAG pipelines with objective metrics.\n- Generate test datasets for evaluation.\n\nFirst <b>action</b>: - Run pip install ragas.\n- Run ragas quickstart rag_eval to create an example project.\n\nPrerequisites","status":"active","categories":["T20"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M6W7JKZ5QX5GMX5XHD3MY2BE","api_url":"https://veyrul.com/v1/records/src_01M6W7JKZ5QX5GMX5XHD3MY2BE","updated_at":"2026-11-08T00:05:20.449Z","score":0.2796}},{"kind":"record","record":{"id":"src_01M6V5B9QQ73F2XFEW9HE7G59T","type":"source","title":"Ragas: toolkit for evaluating LLM applications, README overview","summary":"The README presents Ragas as a toolkit for evaluating and optimizing LLM applications with objective metrics and test data generation.","snippet":"have no data for.\n- Start a RAG evaluation project from a template.\n\nFirst <b>action</b>: - Install with pip install ragas.\n- Run ragas quickstart rag_eval, optionally with -o to pick","status":"active","categories":["T20"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M6V5B9QQ73F2XFEW9HE7G59T","api_url":"https://veyrul.com/v1/records/src_01M6V5B9QQ73F2XFEW9HE7G59T","updated_at":"2026-11-07T14:07:09.006Z","score":0.2794}},{"kind":"record","record":{"id":"src_01M67NXPMVMD6ZP8K6K0ECBGQE","type":"source","title":"Searge-LLM for ComfyUI v1.0 (ComfyUI registry node ComfyUI_Searge_LLM)","summary":"ComfyUI node registry custom node pack ComfyUI_Searge_LLM at revision 3804774923b3: publisher admin-11338bd3-f081-43cf-b3f9-295c829826f7; declared licence {}; repository https://github.com/SeargeDP/ComfyUI_Searge_LLM. Metadata-only index entry; not tested guidance.","snippet":"prompt-improvement node for ComfyUI, utilizing the power of a language model to <b>turn</b> a provided text-to-image prompt into a more detailed and improved prompt.\"\ncreated","status":"active","categories":["T08"],"evidence":{"inspection":"metadata_checked","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M67NXPMVMD6ZP8K6K0ECBGQE","api_url":"https://veyrul.com/v1/records/src_01M67NXPMVMD6ZP8K6K0ECBGQE","updated_at":"2026-10-31T00:32:00.610Z","score":0.2829}},{"kind":"record","record":{"id":"src_01M6VC6PPY5987TWR15Q0MHJ5K","type":"source","title":"LiteLLM: unified AI gateway and Python SDK for 100+ LLM providers, README overview","summary":"The README presents LiteLLM as one OpenAI-style interface over many LLM providers, as an SDK or a gateway.","snippet":"without rewriting code.\n- Run a shared gateway with keys, spend tracking and guardrails.\n\nFirst <b>action</b>: - Install the SDK with uv add litellm.\n- Or deploy the proxy server following the proxy","status":"active","categories":["T09"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M6VC6PPY5987TWR15Q0MHJ5K","api_url":"https://veyrul.com/v1/records/src_01M6VC6PPY5987TWR15Q0MHJ5K","updated_at":"2026-11-07T16:06:58.495Z","score":0.2795}},{"kind":"record","record":{"id":"src_01M6V5B5WDEXSX2S4T4K34DJ3T","type":"source","title":"promptfoo: CLI and library for LLM evals and red teaming, README overview","summary":"The README presents promptfoo as a CLI and library for evaluating and red-teaming LLM-based apps.","snippet":"vulnerabilities with its red teaming mode.\n- Add evaluation checks to CI/CD pipelines.\n\nFirst <b>action</b>: - Install with npm, brew or pip, or run it with npx without installing.\n- Run promptfoo init","status":"active","categories":["T20"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M6V5B5WDEXSX2S4T4K34DJ3T","api_url":"https://veyrul.com/v1/records/src_01M6V5B5WDEXSX2S4T4K34DJ3T","updated_at":"2026-11-07T14:07:05.066Z","score":0.2794}},{"kind":"record","record":{"id":"src_01M6V5KSRQK55AEPZDJT13K257","type":"source","title":"garak: LLM vulnerability scanner, README overview","summary":"The README presents garak as a scanner that checks whether an LLM-backed application can be made to fail, using static, dynamic and adaptive probes.","snippet":"files.\n- List the available probes and select a probe spec for a run.\n\nFirst <b>action</b>: - Install with python -m pip install -U garak, or use a Conda environment","status":"active","categories":["T20"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed"},"url":"https://veyrul.com/sources/src_01M6V5KSRQK55AEPZDJT13K257","api_url":"https://veyrul.com/v1/records/src_01M6V5KSRQK55AEPZDJT13K257","updated_at":"2026-11-07T14:11:47.576Z","score":0.2793}},{"kind":"guide_preview","preview":{"schema":"guide-preview/v1","id":"art_01M6HFSQAGB5MC2H4DYDWNXRKS","type":"article","revision":1,"title":"StoryVerse paper: writing a plot as abstract acts that an LLM planner turns into character actions","coverage":"A source note on the StoryVerse paper for writers and game-narrative designers: how a plot outline written as abstract acts with goals, prerequisites and placeholders is planned into character actions by an LLM. It is based on two paper pages, nothing was run, and the paper reports no evaluation.","example":null,"tips":["Write prerequisites as statements that can be judged true or false against the world state, a player action or an earlier act.","The paper's planner can add characters the acts never named, so read every generated action sequence.","The paper lists user studies and quantitative evaluation as future work, so treat its fable-scale examples as demonstrations only."],"categories":["T20","T01"],"tags":["storyverse","abstract-acts","narrative-planning","character-simulation","dynamic-plot","game-narrative","paper","documented-not-tested"],"evidence":{"inspection":"source_inspected","execution":"not_run","review":"unreviewed","review_revision":null},"media":[]}}],"truncated":true,"has_more":true,"next_cursor":"eyJyZWZlcmVuY2UiOiJNakEiLCJwcmV2aWV3IjpudWxsLCJxIjoiU3RvcnlWZXJzZSBwYXBlcjogd3JpdGluZyBhIHBsb3QgYXMgYWJzdHJhY3QgYWN0cyB0aGF0IGFuIExMTSBwbGFubmVyIHR1cm5zIGludG8gY2hhcmFjdGVyIGFjdGlvbnMiLCJ0eXBlcyI6W10sImNhdGVnb3JpZXMiOltdfQ","ordering":"bounded_reference_and_preview_lanes"}