Recipe Workflow

Overview

The Recipe Workflow provides a complete cycle for running, evaluating, and improving agent configurations.

Run history (Run IDs, statuses, sessions, timings) is captured automatically for every run. The --save flag additionally captures a full replay trace for deep debugging.

Commands

Run a Recipe

praisonai recipe run <recipe-name> [OPTIONS]

Option	Description
`--save`, `-s`	Save replay trace for debugging
`--name`, `-n NAME`	Custom trace name
`--var KEY=VALUE`	Set variables
`--debug`	Enable debug output
`--verbose`, `-v`	Verbose output

Example:

praisonai recipe run ai-topic-gatherer --save --name my-trace --var topic="AI 2026"

Judge a Trace

praisonai recipe judge <trace-id> [OPTIONS]

Option	Description
`--yaml FILE`	YAML for context-aware evaluation
`--output FILE`	Output plan file
`--model MODEL`	Judge model (default: gpt-4o-mini)

Example:

praisonai recipe judge my-trace --yaml agents.yaml --output plan.yaml

Apply Improvements

praisonai recipe apply <plan-file> [OPTIONS]

Option	Description
`--confirm` / `-c`	Apply without confirmation
`--dry-run`	Preview only
`--fix-ids IDS`	Apply specific fixes
`--no-backup`	Skip backup

Example:

# Preview changes
praisonai recipe apply plan.yaml --dry-run

# Apply with confirmation
praisonai recipe apply plan.yaml --confirm

Complete Workflow

Run Recipe with Trace

praisonai recipe run ai-wordpress-post-generator \
  --save --name wp-test \
  --var topic="AI News"

Judge the Execution

Find the run ID with praisonai recipe runs, then pass it to recipe judge.

praisonai recipe judge wp-test \
  --yaml agents.yaml \
  --output improvements.yaml

Review Suggestions

praisonai recipe apply improvements.yaml --dry-run

Apply Improvements

praisonai recipe apply improvements.yaml --confirm

Re-run and Verify

praisonai recipe run ai-wordpress-post-generator \
  --save --name wp-test-v2 \
  --var topic="AI News"

Evaluation Metrics

The judge evaluates each agent on 6 criteria:

Metric	Description
Task Achievement	Did the agent complete its goal?
Context Utilization	Was context effectively used?
Output Quality	Is output high quality?
Instruction Following	Did agent follow instructions?
Hallucination	10=accurate, 1=fabricated
Error Handling	How well were errors handled?

Fix Types

The judge generates actionable fixes:

Type	Description
`append_instruction`	Add guidance to instructions
`add_expected_output`	Specify output format
`modify_context_config`	Adjust context settings
`suggestion`	General improvement

Example Output

Judge Report

============================================================
  LLM JUDGE REPORT: wp-test
============================================================

  Overall Score: 7.86/10
  Agents Evaluated: 6

  AGENT SCORES:
    Research Agent:
      Task Achievement: 8.0/10
      Context Utilization: 7.0/10
      Output Quality: 8.0/10
      Instruction Following: 9.0/10
      Hallucination: 8.0/10
      Error Handling: 10.0/10
      Overall: 8.33/10

  CONTEXT FLOW:
    Research Agent → Writer Agent: 10.0/10 ✓
    Writer Agent → Publisher: 10.0/10 ✓

  RECOMMENDATIONS:
    - Research Agent: Include specific citations
    - Writer Agent: Ensure all sections complete
============================================================

Plan Preview

============================================================
  PREVIEW: Changes to be applied
============================================================

  [MEDIUM] fix_abc123
    Agent: Research Agent
    Type: append_instruction
    Path: agents.Research Agent.instructions
    New: IMPROVEMENT: Include specific citations...

  [LOW] fix_def456
    Agent: Writer Agent
    Type: append_instruction
    New: IMPROVEMENT: Ensure sections complete...

============================================================

Best Practices

History is Automatic

Every run is recorded to history; add --save only for full replay traces

Name Your Traces

Use --name for easy identification

Preview First

Always --dry-run before applying

Iterate

Run → Judge → Apply → Repeat

Viewing Traces

Use the replay commands to inspect traces:

# List all traces
praisonai replay list

# View trace details
praisonai replay context my-trace --dump

# Full content (no truncation)
praisonai replay context my-trace --dump --full

# Statistics
praisonai replay context my-trace --stats

Run History - Auto-recorded runs and the recipe runs command
Context Replay - Trace capture and viewing
LLM as Judge - Evaluation details

Getting Started

Install

Core Concepts

Guides

Features

Models

Databases

Observability

Memory

Telemetry & Analytics

Knowledge

RAG

Persistence

Tools

Other Features

Developers

Configuration

Best Practices

Getting Started (No Code)

Overview

Commands

Run a Recipe

Judge a Trace

Apply Improvements

Complete Workflow

Evaluation Metrics

Fix Types

Example Output

Judge Report

Plan Preview

Best Practices

History is Automatic

Name Your Traces

Preview First

Iterate

Viewing Traces

​Overview

​Commands

​Run a Recipe

​Judge a Trace

​Apply Improvements

​Complete Workflow

​Evaluation Metrics

​Fix Types

​Example Output

​Judge Report

​Plan Preview

​Best Practices

History is Automatic

Name Your Traces

Preview First

Iterate

​Viewing Traces

​Related

Overview

Commands

Run a Recipe

Judge a Trace

Apply Improvements

Complete Workflow

Evaluation Metrics

Fix Types

Example Output

Judge Report

Plan Preview

Best Practices

Viewing Traces

Related