Voice-to-JSON Task Parser

Extract structured to-do items from voice transcripts and meeting notes

POST /api/parse-tasks
AI/ML <150ms
Latency: ~150ms
Auth: API Key
Pricing: Free tier

What This API Does

Transforms unstructured voice transcripts and meeting notes into structured JSON task lists. Detects 60+ action verbs, extracts deadlines, identifies priorities (high/medium/low), and resolves assignees from @mentions and name patterns. Filters out conversational chatter to focus only on actionable items.

Description

The Voice-to-JSON Task Parser uses a multi-layer NLP pipeline built entirely with regular expressions—no machine learning required. This makes it incredibly fast, predictable, and easy to deploy without external dependencies.

Key Features

  • 60+ Action Verbs: Recognizes complete, update, review, schedule, prepare, and many more
  • Smart Date Extraction: Parses natural language dates like "next Friday", "by EOD", "in 2 weeks"
  • Priority Detection: Identifies urgency markers like "urgent", "ASAP", "low priority"
  • Assignee Resolution: Extracts @mentions and name patterns like "John will..." or "Sarah needs to..."
  • Chatter Filtering: Distinguishes actionable tasks from casual conversation

Code Examples

curl -X POST https://api.atomicapis.dev/api/parse-tasks \
  -H "X-RapidAPI-Proxy-Secret: YOUR_SECRET" \
  -H "Content-Type: application/json" \
  -d '{
    "transcript": "Okay team, @Sarah needs to complete the Q4 report by Friday. John should review the design mockups ASAP. Mike, can you prepare the slides by end of day?",
    "defaultAssignee": null,
    "confidenceThreshold": 0.3
  }'

Request Parameters

Parameter Type Required Description
transcript string Required The voice transcript or meeting notes text to parse. Maximum 100,000 characters.
defaultAssignee string Optional Fallback assignee name when no assignee is detected for a task.
confidenceThreshold double Optional Minimum confidence score (0.0-1.0) to include a task. Default: 0.3

Response Format

Example Response (200 OK)
{
  "transcriptLength": 168,
  "sentencesAnalyzed": 3,
  "tasksDetected": 3,
  "tasks": [
    {
      "taskNumber": 1,
      "description": "Complete the Q4 report",
      "rawSentence": "@Sarah needs to complete the Q4 report by Friday.",
      "assignee": "Sarah",
      "priority": "medium",
      "deadline": "Friday",
      "confidence": 0.85,
      "actionVerb": "complete"
    },
    {
      "taskNumber": 2,
      "description": "Review the design mockups",
      "rawSentence": "John should review the design mockups ASAP.",
      "assignee": "John",
      "priority": "high",
      "deadline": null,
      "confidence": 0.82,
      "actionVerb": "review"
    },
    {
      "taskNumber": 3,
      "description": "Prepare the slides",
      "rawSentence": "Mike, can you prepare the slides by end of day?",
      "assignee": "Mike",
      "priority": "medium",
      "deadline": "end of day",
      "confidence": 0.78,
      "actionVerb": "prepare"
    }
  ],
  "parseDurationMs": 12.5
}

Response Fields

Field Type Description
transcriptLength int Character count of the input transcript
sentencesAnalyzed int Number of sentences analyzed
tasksDetected int Number of tasks extracted above threshold
tasks[].taskNumber int Sequential task number
tasks[].description string Cleaned task description
tasks[].rawSentence string Original sentence the task was extracted from
tasks[].assignee string | null Detected assignee name or null
tasks[].priority string Priority level: high, medium, or low
tasks[].deadline string | null Detected deadline text or null
tasks[].confidence double Confidence score from 0.0 to 1.0
tasks[].actionVerb string | null Primary action verb identified
parseDurationMs double Processing time in milliseconds

Use Cases

Meeting Notes

Convert meeting transcripts into actionable task lists automatically. Never miss a follow-up action item again.

// Perfect for standups, retros, planning

Voice Memos

Dictate tasks on the go and instantly convert them to structured data. Ideal for mobile productivity workflows.

// Dictate → Parse → Sync to your tools

Project Management

Integrate with project management tools to auto-create tickets from Slack messages, emails, or team discussions.

// Auto-create Jira/Trello/Asana cards

Build Constraints

Pure Regex NLP Pipeline

This API is built entirely using regular expressions—no machine learning models, no external NLP services, no GPU requirements. This design choice ensures:

Predictable Performance

Consistent ~150ms response times regardless of input complexity

Zero External Dependencies

No ML model servers, no third-party APIs, no rate limit surprises

Horizontal Scaling

Stateless processing allows infinite horizontal scaling

Fully Explainable

Every extraction decision can be traced to a specific regex pattern

NLP Pipeline Layers

1
Action Verb Detection

Pattern matching against 60+ predefined action verbs

2
Modal/Imperative Pattern Matching

Identifies sentences using modal verbs and imperative structures

3
Date Extraction

Parses relative dates, absolute dates, and time expressions

4
Assignee Resolution

Extracts @mentions and name patterns from sentence context

Error Codes

Code Status Description
400 Bad Request Missing or invalid transcript parameter
401 Unauthorized Invalid or missing API key
413 Payload Too Large Transcript exceeds 100,000 character limit
429 Rate Limited Too many requests, retry after the specified time
500 Server Error Internal processing error, please try again

Ready to Parse Your Tasks?

Start extracting structured tasks from voice transcripts and meeting notes in minutes.

MCP Integration MCP Ready

What is MCP?

Model Context Protocol (MCP) allows AI assistants like Claude to call this API as a native tool during conversation. Instead of writing HTTP requests, the AI invokes the tool directly — no API keys or boilerplate needed on the client side.

Tool Details

Tool Class
VoiceTaskParserTools
Method
ParseTasks()

Description

Extracts actionable tasks from voice transcripts with assignees and deadlines