CSV, JSON, JSONL or Parquet — your call. Same e-commerce schema in all four, no conversion step. Column order and signal fields stay identical. https://www.socialintel.io/custom-dataset?utm_source=social&utm_medium=custom_dataset&utm_channel=bluesky&utm_campaign=ecommerce
November 10, 2026 at 2:00 PM
CSV, JSON, JSONL or Parquet — your call. Same e-commerce schema in all four, no conversion step. Column order and signal fields stay identical. https://www.socialintel.io/custom-dataset?utm_source=social&utm_medium=custom_dataset&utm_channel=bluesky&utm_campaign=ecommerce
OpenRouter, OpenAI & Anthropic Usage CSV Bill Analyzer: 100% Client-Side LLM FinOps Audit Calculator
Drop your 30-day OpenRouter, Anthropic Console, OpenAI API, or Cursor/Cline usage CSV or JSONL file to run a 100% browser-local FinOps audit. Instantly detect uncached prefix waste, multi-turn ...
Drop your 30-day OpenRouter, Anthropic Console, OpenAI API, or Cursor/Cline usage CSV or JSONL file to run a 100% browser-local FinOps audit. Instantly detect uncached prefix waste, multi-turn ...
November 10, 2026 at 11:15 AM
OpenRouter, OpenAI & Anthropic Usage CSV Bill Analyzer: 100% Client-Side LLM FinOps Audit Calculator
Drop your 30-day OpenRouter, Anthropic Console, OpenAI API, or Cursor/Cline usage CSV or JSONL file to run a 100% browser-local FinOps audit. Instantly detect uncached prefix waste, multi-turn ...
Drop your 30-day OpenRouter, Anthropic Console, OpenAI API, or Cursor/Cline usage CSV or JSONL file to run a 100% browser-local FinOps audit. Instantly detect uncached prefix waste, multi-turn ...
That's right but also intentional
Storing all the data in spaces solves a lot of the network load issue. I'm also thinking we'd store the actual events as a blob (zipped jsonl) and batch up writes to avoid overly straining the system.
The big win on having records in a PDS is user deletion.
Storing all the data in spaces solves a lot of the network load issue. I'm also thinking we'd store the actual events as a blob (zipped jsonl) and batch up writes to avoid overly straining the system.
The big win on having records in a PDS is user deletion.
November 10, 2026 at 1:46 AM
That's right but also intentional
Storing all the data in spaces solves a lot of the network load issue. I'm also thinking we'd store the actual events as a blob (zipped jsonl) and batch up writes to avoid overly straining the system.
The big win on having records in a PDS is user deletion.
Storing all the data in spaces solves a lot of the network load issue. I'm also thinking we'd store the actual events as a blob (zipped jsonl) and batch up writes to avoid overly straining the system.
The big win on having records in a PDS is user deletion.
自動化に組み込んでから、出力の形が違うと気づいて、全部書き直し。
私はこれをやった。
片方は出力形式を選べて、もう片方はイベントを JSONL で流す。
Claude Code と Codex CLI。この差の吸収は、先に決める。
#ClaudeCode #CodexCLI
私はこれをやった。
片方は出力形式を選べて、もう片方はイベントを JSONL で流す。
Claude Code と Codex CLI。この差の吸収は、先に決める。
#ClaudeCode #CodexCLI
November 8, 2026 at 5:20 AM
自動化に組み込んでから、出力の形が違うと気づいて、全部書き直し。
私はこれをやった。
片方は出力形式を選べて、もう片方はイベントを JSONL で流す。
Claude Code と Codex CLI。この差の吸収は、先に決める。
#ClaudeCode #CodexCLI
私はこれをやった。
片方は出力形式を選べて、もう片方はイベントを JSONL で流す。
Claude Code と Codex CLI。この差の吸収は、先に決める。
#ClaudeCode #CodexCLI
CSV, JSON, JSONL or Parquet — your call. Same media schema in all four. https://www.socialintel.io/custom-dataset?utm_source=social&utm_medium=custom_dataset&utm_channel=bluesky&utm_campaign=media
November 10, 2026 at 2:00 PM
CSV, JSON, JSONL or Parquet — your call. Same media schema in all four. https://www.socialintel.io/custom-dataset?utm_source=social&utm_medium=custom_dataset&utm_channel=bluesky&utm_campaign=media
flux3_ai_prompt_tools is a handy offline ruby library for validating and assembling structured prompts into json/jsonl manifests for flux image generation, no api calls needed.
https://rubycoder.ai/item/flux3-ai-prompt-tools
#ruby #devtools
https://rubycoder.ai/item/flux3-ai-prompt-tools
#ruby #devtools
flux3_ai_prompt_tools
An offline Ruby library and CLI tool that validates and assembles structured prompts into JSON and JSONL manifests for text-to-image generation, without calling APIs or sending data over the network.
rubycoder.ai
November 9, 2026 at 8:00 AM
flux3_ai_prompt_tools is a handy offline ruby library for validating and assembling structured prompts into json/jsonl manifests for flux image generation, no api calls needed.
https://rubycoder.ai/item/flux3-ai-prompt-tools
#ruby #devtools
https://rubycoder.ai/item/flux3-ai-prompt-tools
#ruby #devtools
Hate converting files after download? CSV, JSON, JSONL, and Parquet — same columns, no migration script. https://www.socialintel.io/custom-dataset?utm_source=social&utm_medium=custom_dataset&utm_channel=bluesky&utm_campaign=gaming
November 10, 2026 at 6:38 AM
Hate converting files after download? CSV, JSON, JSONL, and Parquet — same columns, no migration script. https://www.socialintel.io/custom-dataset?utm_source=social&utm_medium=custom_dataset&utm_channel=bluesky&utm_campaign=gaming
📦 koh/sylius-acp v1.3.5
Agentic Commerce Protocol for Sylius: ACP carts with a buyer handoff, /.well-known/acp.json, and the ACP product feed published to agents (JSONL snapshot or Feed API upsert).
🔗 https://github.com/ohartwig/sylius-acp
Agentic Commerce Protocol for Sylius: ACP carts with a buyer handoff, /.well-known/acp.json, and the ACP product feed published to agents (JSONL snapshot or Feed API upsert).
🔗 https://github.com/ohartwig/sylius-acp
November 11, 2026 at 9:09 PM
📦 koh/sylius-acp v1.3.5
Agentic Commerce Protocol for Sylius: ACP carts with a buyer handoff, /.well-known/acp.json, and the ACP product feed published to agents (JSONL snapshot or Feed API upsert).
🔗 https://github.com/ohartwig/sylius-acp
Agentic Commerce Protocol for Sylius: ACP carts with a buyer handoff, /.well-known/acp.json, and the ACP product feed published to agents (JSONL snapshot or Feed API upsert).
🔗 https://github.com/ohartwig/sylius-acp
Four AI coding harnesses, one laptop, one task: they now dispatch work to each other. One drafts, one verifies, failures get logged and reviewed for patterns. Multi-agent was a slide deck two years ago. Today it is a shell script and a JSONL file.
November 11, 2026 at 9:00 PM
Four AI coding harnesses, one laptop, one task: they now dispatch work to each other. One drafts, one verifies, failures get logged and reviewed for patterns. Multi-agent was a slide deck two years ago. Today it is a shell script and a JSONL file.
Docker Sandboxes Audit Logs: Reading Local JSONL Records and Querying the AI Governance Audit Logs A
A hands-on guide to AI Governance Audit Logs for Docker Sandboxes: confir...
https://collabnix.com/docker-sandboxes-audit-logs-local-jsonl-audit-logs-api/
#Docker #AI
A hands-on guide to AI Governance Audit Logs for Docker Sandboxes: confir...
https://collabnix.com/docker-sandboxes-audit-logs-local-jsonl-audit-logs-api/
#Docker #AI
November 7, 2026 at 11:27 AM
Docker Sandboxes Audit Logs: Reading Local JSONL Records and Querying the AI Governance Audit Logs A
A hands-on guide to AI Governance Audit Logs for Docker Sandboxes: confir...
https://collabnix.com/docker-sandboxes-audit-logs-local-jsonl-audit-logs-api/
#Docker #AI
A hands-on guide to AI Governance Audit Logs for Docker Sandboxes: confir...
https://collabnix.com/docker-sandboxes-audit-logs-local-jsonl-audit-logs-api/
#Docker #AI
I hear you but it's the era of effortless datalakes so file format is ultimately less important. You can build a smart sql query capable tool on top of JSONL, parquet, whatever. Using sqlalchemy just makes creating that tool a lot easier imo
Reason no litellm is at the top of the list. Using jsonl instead of SQL is more of a personal preference since polars and pandas can treat most file formats almost interchangeably. Config management works quietly in background unappreciated until scope creep makes it a foot gun.
November 11, 2026 at 7:02 PM
I hear you but it's the era of effortless datalakes so file format is ultimately less important. You can build a smart sql query capable tool on top of JSONL, parquet, whatever. Using sqlalchemy just makes creating that tool a lot easier imo
Four AI coding harnesses, one laptop, one task: they now dispatch work to each other. One drafts, one verifies, failures get logged and reviewed for patterns. Multi-agent was a slide deck two years ago. Today it is a shell script and a JSONL file.
November 7, 2026 at 9:00 PM
Four AI coding harnesses, one laptop, one task: they now dispatch work to each other. One drafts, one verifies, failures get logged and reviewed for patterns. Multi-agent was a slide deck two years ago. Today it is a shell script and a JSONL file.
Four AI coding harnesses, one laptop, one task: they now dispatch work to each other. One drafts, one verifies, failures get logged and reviewed for patterns. Multi-agent was a slide deck two years ago. Today it is a shell script and a JSONL file.
November 10, 2026 at 9:00 PM
Four AI coding harnesses, one laptop, one task: they now dispatch work to each other. One drafts, one verifies, failures get logged and reviewed for patterns. Multi-agent was a slide deck two years ago. Today it is a shell script and a JSONL file.
OpenRouter, OpenAI & Anthropic Usage CSV Bill Analyzer: 100% Client-Side LLM FinOps Audit Calculator
Drop your 30-day OpenRouter, Anthropic Console, OpenAI API, or Cursor/Cline usage CSV or JSONL file to run a 100% browser-local FinOps audit. Instantly detect uncached prefix waste, multi-turn ...
Drop your 30-day OpenRouter, Anthropic Console, OpenAI API, or Cursor/Cline usage CSV or JSONL file to run a 100% browser-local FinOps audit. Instantly detect uncached prefix waste, multi-turn ...
November 11, 2026 at 1:15 AM
OpenRouter, OpenAI & Anthropic Usage CSV Bill Analyzer: 100% Client-Side LLM FinOps Audit Calculator
Drop your 30-day OpenRouter, Anthropic Console, OpenAI API, or Cursor/Cline usage CSV or JSONL file to run a 100% browser-local FinOps audit. Instantly detect uncached prefix waste, multi-turn ...
Drop your 30-day OpenRouter, Anthropic Console, OpenAI API, or Cursor/Cline usage CSV or JSONL file to run a 100% browser-local FinOps audit. Instantly detect uncached prefix waste, multi-turn ...
Storyblok’s CLI now has a find command: search stories, filter with JSONPath, and pipe JSONL output into jq or your own scripts.
A few flags took one query from 603s to 1.8s. 👀
www.storyblok.com/mp/find-ever...
A few flags took one query from 603s to 1.8s. 👀
www.storyblok.com/mp/find-ever...
November 9, 2026 at 9:13 AM
Storyblok’s CLI now has a find command: search stories, filter with JSONPath, and pipe JSONL output into jq or your own scripts.
A few flags took one query from 603s to 1.8s. 👀
www.storyblok.com/mp/find-ever...
A few flags took one query from 603s to 1.8s. 👀
www.storyblok.com/mp/find-ever...
Four agents and one JSONL file, that's not a pipeline, that's a tiny startup. My advice: log the prompts too, not just the failures, or you'll debug ghosts.
November 7, 2026 at 10:42 PM
Four agents and one JSONL file, that's not a pipeline, that's a tiny startup. My advice: log the prompts too, not just the failures, or you'll debug ghosts.
OpenRouter, OpenAI & Anthropic Usage CSV Bill Analyzer: 100% Client-Side LLM FinOps Audit Calculator
Drop your 30-day OpenRouter, Anthropic Console, OpenAI API, or Cursor/Cline usage CSV or JSONL file to run a 100% browser-local FinOps audit. Instantly detect uncached prefix waste, multi-turn ...
Drop your 30-day OpenRouter, Anthropic Console, OpenAI API, or Cursor/Cline usage CSV or JSONL file to run a 100% browser-local FinOps audit. Instantly detect uncached prefix waste, multi-turn ...
November 10, 2026 at 1:15 AM
OpenRouter, OpenAI & Anthropic Usage CSV Bill Analyzer: 100% Client-Side LLM FinOps Audit Calculator
Drop your 30-day OpenRouter, Anthropic Console, OpenAI API, or Cursor/Cline usage CSV or JSONL file to run a 100% browser-local FinOps audit. Instantly detect uncached prefix waste, multi-turn ...
Drop your 30-day OpenRouter, Anthropic Console, OpenAI API, or Cursor/Cline usage CSV or JSONL file to run a 100% browser-local FinOps audit. Instantly detect uncached prefix waste, multi-turn ...
Reason no litellm is at the top of the list. Using jsonl instead of SQL is more of a personal preference since polars and pandas can treat most file formats almost interchangeably. Config management works quietly in background unappreciated until scope creep makes it a foot gun.
November 9, 2026 at 9:19 PM
Reason no litellm is at the top of the list. Using jsonl instead of SQL is more of a personal preference since polars and pandas can treat most file formats almost interchangeably. Config management works quietly in background unappreciated until scope creep makes it a foot gun.
エージェントの記憶に、OpenTelemetryのような共通仕様は必要か
AIエージェント向けの「memory」が、いまかなりバラバラな形で広がっている。あるものはMarkdownファイル、あるものはJSONLの知識グラフ、別のもの…
#HackerNews #AI #テック
AIエージェント向けの「memory」が、いまかなりバラバラな形で広がっている。あるものはMarkdownファイル、あるものはJSONLの知識グラフ、別のもの…
#HackerNews #AI #テック
エージェントの記憶に、OpenTelemetryのような共通仕様は必要か
AIエージェント向けの「memory」が、いまかなりバラバラな形で広がっている。あるものはMarkdownファイル、あるものはJSONLの知識グラフ、別のもの…
papoo.work
November 9, 2026 at 12:08 AM
エージェントの記憶に、OpenTelemetryのような共通仕様は必要か
AIエージェント向けの「memory」が、いまかなりバラバラな形で広がっている。あるものはMarkdownファイル、あるものはJSONLの知識グラフ、別のもの…
#HackerNews #AI #テック
AIエージェント向けの「memory」が、いまかなりバラバラな形で広がっている。あるものはMarkdownファイル、あるものはJSONLの知識グラフ、別のもの…
#HackerNews #AI #テック
Four AI coding harnesses, one laptop, one task: they now dispatch work to each other. One drafts, one verifies, failures get logged and reviewed for patterns. Multi-agent was a slide deck two years ago. Today it is a shell script and a JSONL file.
November 6, 2026 at 9:00 PM
Four AI coding harnesses, one laptop, one task: they now dispatch work to each other. One drafts, one verifies, failures get logged and reviewed for patterns. Multi-agent was a slide deck two years ago. Today it is a shell script and a JSONL file.
OpenRouter, OpenAI & Anthropic Usage CSV Bill Analyzer: 100% Client-Side LLM FinOps Audit Calculator
Drop your 30-day OpenRouter, Anthropic Console, OpenAI API, or Cursor/Cline usage CSV or JSONL file to run a 100% browser-local FinOps audit. Instantly detect uncached prefix waste, multi-turn ...
Drop your 30-day OpenRouter, Anthropic Console, OpenAI API, or Cursor/Cline usage CSV or JSONL file to run a 100% browser-local FinOps audit. Instantly detect uncached prefix waste, multi-turn ...
November 9, 2026 at 2:45 AM
OpenRouter, OpenAI & Anthropic Usage CSV Bill Analyzer: 100% Client-Side LLM FinOps Audit Calculator
Drop your 30-day OpenRouter, Anthropic Console, OpenAI API, or Cursor/Cline usage CSV or JSONL file to run a 100% browser-local FinOps audit. Instantly detect uncached prefix waste, multi-turn ...
Drop your 30-day OpenRouter, Anthropic Console, OpenAI API, or Cursor/Cline usage CSV or JSONL file to run a 100% browser-local FinOps audit. Instantly detect uncached prefix waste, multi-turn ...
Four AI coding harnesses, one laptop, one task: they now dispatch work to each other. One drafts, one verifies, failures get logged and reviewed for patterns. Multi-agent was a slide deck two years ago. Today it is a shell script and a JSONL file.
November 8, 2026 at 9:00 PM
Four AI coding harnesses, one laptop, one task: they now dispatch work to each other. One drafts, one verifies, failures get logged and reviewed for patterns. Multi-agent was a slide deck two years ago. Today it is a shell script and a JSONL file.
PythonでJSONLの壊れた行番号を調べる：データを表示しない読み取り専用チェッカー
Python標準ライブラリでJSONLの壊れた行、UTF-8、BOM、空行、NaNを検出。実測13テストと終了コード、機密データを表示しない設計を解説。
Python標準ライブラリでJSONLの壊れた行、UTF-8、BOM、空行、NaNを検出。実測13テストと終了コード、機密データを表示しない設計を解説。
PythonでJSONLの壊れた行番号を調べる：データを表示しない読み取り専用チェッカー
Python標準ライブラリでJSONLの壊れた行、UTF-8、BOM、空行、NaNを検出。実測13テストと終了コード、機密データを表示しない設計を解説。
papanda925.com
November 5, 2026 at 9:05 PM
PythonでJSONLの壊れた行番号を調べる：データを表示しない読み取り専用チェッカー
Python標準ライブラリでJSONLの壊れた行、UTF-8、BOM、空行、NaNを検出。実測13テストと終了コード、機密データを表示しない設計を解説。
Python標準ライブラリでJSONLの壊れた行、UTF-8、BOM、空行、NaNを検出。実測13テストと終了コード、機密データを表示しない設計を解説。
📦 koh/sylius-acp v1.3.4
Agentic Commerce Protocol for Sylius: ACP carts with a buyer handoff, /.well-known/acp.json, and the ACP product feed published to agents (JSONL snapshot or Feed API upsert).
🔗 https://github.com/ohartwig/sylius-acp
Agentic Commerce Protocol for Sylius: ACP carts with a buyer handoff, /.well-known/acp.json, and the ACP product feed published to agents (JSONL snapshot or Feed API upsert).
🔗 https://github.com/ohartwig/sylius-acp
November 8, 2026 at 1:00 AM
📦 koh/sylius-acp v1.3.4
Agentic Commerce Protocol for Sylius: ACP carts with a buyer handoff, /.well-known/acp.json, and the ACP product feed published to agents (JSONL snapshot or Feed API upsert).
🔗 https://github.com/ohartwig/sylius-acp
Agentic Commerce Protocol for Sylius: ACP carts with a buyer handoff, /.well-known/acp.json, and the ACP product feed published to agents (JSONL snapshot or Feed API upsert).
🔗 https://github.com/ohartwig/sylius-acp