> For the complete documentation index, see [llms.txt](https://www.pranaypourkar.co.in/the-programmers-guide/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://www.pranaypourkar.co.in/the-programmers-guide/ai/generative-ai/large-language-models-llm/prompt-engineering/prompt-engineering-techniques/4.-output-control-techniques/constrained-formatting.md).

# Constrained formatting

## About

Constrained Formatting is a technique where the model is instructed to **generate output in a strictly controlled format**, using predefined patterns such as:

* Bullet points
* Numbered lists
* Tables
* Fixed templates
* Key-value pairs
* Delimited blocks

Unlike general structured output (which focuses on readability), constrained formatting focuses on:

* **Consistency**
* **Strict adherence to format**
* **Limited variation**

Core idea:

> Restrict how the output is written, not just what it contains.

This ensures that outputs are:

* Predictable
* Uniform across runs
* Easy to scan or parse

## Why Constrained Formatting Is Critical ?

Without constraints, outputs may:

* Change structure between runs
* Mix paragraphs and lists
* Include unnecessary explanations
* Be hard to parse or compare

In engineering workflows, this causes:

* Parsing issues
* Inconsistent UI rendering
* Difficulty in automation
* Poor readability

Constrained formatting ensures:

* Consistent structure
* Reduced variability
* Easier integration with systems
* Better human readability

## The Purpose of Constrained Formatting

This technique aims to:

1. Enforce consistent output layout
2. Improve readability and scan-ability
3. Reduce structural variability
4. Support partial automation (parsing lists/tables)
5. Align outputs with UI or reporting formats

It transforms output from:

Flexible text → Controlled format

## Where Constrained Formatting Fits in the Prompt Lifecycle ?

```
Problem Definition
      ↓
Input Design
      ↓
Reasoning
      ↓
Knowledge Grounding
      ↓
Formatting Constraints  ← (Apply strict output structure)
      ↓
Final Output
```

It acts as a **format enforcement layer before delivery**.

## Different Constrained Formatting Patterns

#### 1. Fixed Bullet Format

Define exact number and structure:

* Point 1:
* Point 2:
* Point 3:

Instruction example:

“Provide exactly 3 bullet points. Each point must be one sentence.”

#### 2. Numbered Step Format

Useful for processes:

1. Step 1
2. Step 2
3. Step 3

Ensures ordered and consistent output.

#### 3. Table Format

Define column structure:

| Field | Description |
| ----- | ----------- |

Useful for:

* Comparisons
* Validation reports
* Structured summaries

#### 4. Key-Value Format

Strict pairing:

Field: Value\
Field: Value

Useful for:

* Logs
* Metadata
* Configurations

#### 5. Delimited Blocks

Use markers:

START\
...\
END

Prevents mixing of content and improves parsing.

## Common Mistakes

#### 1. Not Defining Exact Constraints

Weak:\
“Use bullet points.”

Strong:\
“Provide exactly 4 bullet points, each under 15 words.”

Precision reduces variability.

#### 2. Allowing Extra Text

If not restricted:

* Model may add explanations before/after format

Always specify:

“Do not include text outside the defined format.”

#### 3. Mixing Multiple Formats

Example:

* Bullet points + paragraphs + tables

Leads to inconsistency and confusion.

Stick to one format per response.

#### 4. Not Defining Limits

If count is not specified:

* Number of items may vary

Always define:

* Number of bullets
* Number of rows
* Length constraints

#### 5. Ignoring Edge Cases

If output is empty or partial:

* Format may break

Define fallback:

“If no data, return empty list in same format.”

## Sample Prompts

### Without Constrained Formatting

```
List the benefits of microservices architecture.
```

Issues:

* Variable number of points
* Mixed formatting
* Inconsistent structure

### With Constrained Formatting

```
List the benefits of microservices architecture.

Provide exactly 4 bullet points.
Each bullet must be one concise sentence.
Do not include any text outside the bullet list.
```

Benefits:

* Consistent output
* Predictable structure
* Easy to scan and compare
* Reduced variability


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://www.pranaypourkar.co.in/the-programmers-guide/ai/generative-ai/large-language-models-llm/prompt-engineering/prompt-engineering-techniques/4.-output-control-techniques/constrained-formatting.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.