JSONL
JSON Lines (JSONL) is a line-oriented text format in which each line contains one complete, independently parseable JSON value.
AI Agents for Offensive Security
A newline-delimited JSON format in which each line contains one complete JSON object.
- How it is used
- Records agent observations as appendable artifacts for repeatable reconnaissance pipelines, audit trails, and downstream reasoning.
- How deep it goes
- Defines line-delimited records and implements schemas, append writers, timestamps, streaming, and safety controls without specifying a universal JSONL standard.
- Examples
- Demonstrates a Python artifact logger, a versioned reconnaissance record, consecutive host records, and jq or Python IP-clustering analysis.
- What you can do after
- Operate/configure append-only JSONL artifact logging and safely prepare structured reconnaissance records for review, debugging, and AI analysis.
Data Storytelling with Altair and AI
JSONL is a line-oriented text format in which each line contains a separate JSON value, commonly an object.
- How it is used
- Prepares prompt–completion samples for uploading and fine-tuning a chat model.
- How deep it goes
- Shows the messages schema and a pandas-based CSV conversion script, without detailing JSONL syntax beyond the training structure.
- Examples
- Uses household chores, data-commentary prompts, and CSV rows converted into messages records.
- What you can do after
- Apply it via Python and the OpenAI API to prepare and upload a fine-tuning dataset.
Read how Data Storytelling with Altair and AI explains JSONL →
Logs and Telemetry
JSON Lines is a serialization format that represents each JSON value as a separate line.
- How it is used
- Formats Fluent Bit console output so each log event appears as a separate JSON line.
- How deep it goes
- Explains output configuration, record separation, tag matching, and timestamp formatting, but not JSON parsing or downstream consumption.
- Examples
- Uses dummy inputs with two stdout outputs and an HTTP input configuration that prints records to the console.
- What you can do after
- Operate/configure it by setting Fluent Bit stdout format=json_lines and optional ISO 8601 received-date fields.
DuckDB in Action
JSON Lines is a text format that stores one complete JSON value, typically an object, per line.
- How it is used
- Uses line-oriented JSON files as an immediately queryable source for exploratory analytics.
- How deep it goes
- Explains file discovery, schema inference, array unnesting, recursive expansion, and whole-STRUCT casting.
- Examples
- Demonstrates Premier League shot files, wildcard reads, home-and-away arrays, inferred schemas, and expected-goals aggregation.
- What you can do after
- Apply it via DuckDB SQL to inspect, flatten, type-correct, and aggregate multiple JSONL files.
Also covered in
- LLM Customization and Fine-TuningIn early access (MEAP)
- Data Analysis with LLMs
- Microsoft Azure in Action
- Machine Learning for Tabular Data: XGBoost, Deep Learning, and AI
- Terraform in Depth: Infrastructure as Code with Terraform and OpenTofu
What to know first
- JSONJSON is a language-independent text-based data-interchange format that represents structured values using objects, arrays, strings, numbers, booleans, and null.
- GPT modelA GPT model is a generative, autoregressive Transformer language model in the family developed by OpenAI.
- JSON objectA JSON object is an unordered collection of string-name and JSON-value pairs, represented within curly braces.
- pretrained modelA pretrained model is a machine-learning model whose parameters were learned through prior training before it is used for inference or adapted to a later task.
Related concepts
- AI agentuses · An AI agent is a computational system that uses artificial intelligence to interpret information and select or perform actions toward goals while interacting with an environment.
- fine-tuningused by · Fine-tuning is the continued training of a pretrained model on targeted data or objectives to adapt its parameters or behavior to a task, domain, or use.
- JSON objectuses · A JSON object is an unordered collection of string-name and JSON-value pairs, represented within curly braces.
- SFTTrainerused by · SFTTrainer is a trainer component in the TRL library for configuring and running supervised fine-tuning of language models.
- transcriptused by · A transcript is an ordered written or machine-readable record of content or events from a communication, interaction, or session.
Questions about JSONL
What is JSONL?
JSON Lines (JSONL) is a line-oriented text format in which each line contains one complete, independently parseable JSON value.
Which Manning books explain JSONL?
AI Agents for Offensive Security, Data Storytelling with Altair and AI, Logs and Telemetry and DuckDB in Action.
What should I know before learning JSONL?
Manning books that teach JSONL usually introduce JSON, GPT model, JSON object and pretrained model first.
What is JSONL also called?
JSONL is also called JSON Lines.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.



