Github

phospho

Guides

API Reference

Examples

Go to phospho

Contact us

Run event detection

Created At

Notes

Org Id

Project Id

Session Id

Source

Task Id

Test Id

Value

Eval

Removed

System Prompt

EvaluationModel

Confirmed

Event Name

Messages

Removal Reason

Webhook

Event

Description

Detection Engine

Detection Scope

Is Last Task

Keywords

Recipe Id

Recipe Type

Regex Pattern

Webhook Headers

EventDefinition

Flag

HumanEval

Content

Role

Message

Previous Messages

Corrected Label

Corrected Value

Label

Options Confidence

Score Type

ScoreRange

ScoreRangeSettings

Magnitude

Score

SentimentObject

Additional Input

Additional Output

Data

Environment

Events

Input

Language

Output

Task Position

Topics

Task

Location

Error Type

ValidationError

Run the main pipeline of a project on a task. Returns the events, evals, sentiment, and every other analytics enabled for the phospho project.

The Authorization header must be set with an API key created on the phospho platform. This API key must have billing enabled with a valid payment method.

RunPipelineOnMessagesRequest

HTTPBearer

The results of a pipeline execution. Contains every category of analytics.

Each category of analytics is a dict linking the message_id or task_id
to the analytics object. This is useful for batching and unordered processing.

PipelineResults

HTTPValidationError

Run Main Pipeline On Messages

phospho is the open source text analytics platform for LLM apps

Welcome to phospho!

Clusterize your text messages in 5 minutes. No code required.

Getting started

You can upload your data to phospho by importing a CSV file or an Excel file

Import a CSV or Excel file

Log text messages to phospho in real time

Setup logging in your app

Import from Langsmith 🦜🔗

Import from Langfuse 🪢

Self-hosting

Use the Phospho CLI to interact with the Phospho API

Command Line Interface

Run AB tests in your app to see which version performs better

AB Testing

Events are the key to getting insights from your data in phospho

Event detection

phospho automatically tags your data and warns you when needed

Automatic tagging

Group users messages based on their intention

Clustering

Group tasks together into sessions and attach them to users.

Sessions and Users

Language Detection

Sentiment Analysis

How to send the user feedback from your app to phospho?

User Feedback

Log to phospho with Python

Run custom analytics jobs in Python on your phospho logs

Analyze your logs in Python

Test your agent before deploying it to production

Testing with Python

Log to phospho with Javascript

Add AI analytics to your Langchain agent with phospho

Log to phospho in Python Langchain

Add AI analytics to your Supabase chatbot with phospho

Log to phospho in a Supabase app with a webhook

Export your data to Argilla

Export your data to a PostgreSQL database

Export your data to PostgreSQL

Export your data to create a PowerBI report

Export your data to PowerBI

Open Source Text Analytics Platform for LLM apps

Welcome!

Discover how to run analytics on your LLM app data in 5 minutes

Run analytics on your LLM app data

Learn how to setup LLM as a judge on the phospho platform

LLM as a Judge

Cluster User Intentions

Learn how to visualize, clusterize, and more.

Understand your data

You can generate a dataset from your project and export it to Argilla in one click.

Export a dataset to Argilla

Start logging your fist text message using phospho API

Log messages

Evaluate new system prompt on historical data

Backtesting

Easily query analytics from your phospho project

Query your analytics

Implement logging and discover phospho with those examples

Phospho examples

Generate specific embeddings with phospho

API Reference

Log

Run

Export

Run event detection

Authorizations

Body

Response