Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/better-auth/better-hub/llms.txt

Use this file to discover all available pages before exploring further.

Better Hub provides AI chat capabilities for interacting with GitHub repositories, pull requests, issues, and general GitHub operations.

​
POST /api/ai/ghost

Create or continue an AI conversation with context-aware GitHub assistance. Supports streaming responses with tool execution.
​
messages
array
required
Array of conversation messages with role and content
{
  role: 'user' | 'assistant' | 'system',
  content: string,
  parts?: Array<{type: 'text', text: string}>
}
​
prContext
object
Pull request context for PR-specific assistance
​
owner
string
required
Repository owner
​
repo
string
required
Repository name
​
pullNumber
number
required
Pull request number
​
prTitle
string
required
Pull request title
​
prBody
string
Pull request description
​
baseBranch
string
required
Base branch name
​
headBranch
string
required
Head branch name
​
files
array
Array of changed files with patches
​
mergeConflict
boolean
Whether the PR has merge conflicts
​
issueContext
object
Issue context for issue-specific assistance
​
owner
string
required
Repository owner
​
repo
string
required
Repository name
​
issueNumber
number
required
Issue number
​
title
string
required
Issue title
​
body
string
Issue description
​
state
string
Issue state (open/closed)
​
labels
array
Array of label names
​
pageContext
object
Current page context for better assistance
​
pathname
string
Current page pathname

​
Response (Streaming)

Returns a streaming response with AI-generated text and tool execution results.
​
stream
ReadableStream
Server-sent events stream containing:
  • Text chunks from the AI model
  • Tool execution results
  • Usage statistics
const response = await fetch('/api/ai/ghost', {
  method: 'POST',
  headers: { 'Content-Type': 'application/json' },
  body: JSON.stringify({
    messages: [
      { role: 'user', content: 'Review this pull request for potential issues' }
    ],
    prContext: {
      owner: 'acme',
      repo: 'app',
      pullNumber: 42,
      prTitle: 'Add user authentication',
      prBody: 'Implements JWT-based authentication',
      baseBranch: 'main',
      headBranch: 'feature/auth',
      files: [
        { filename: 'auth.ts', patch: '...' }
      ]
    }
  })
});

const reader = response.body.getReader();
while (true) {
  const { done, value } = await reader.read();
  if (done) break;
  // Process streaming chunks
}
This endpoint consumes AI credits and is subject to rate limiting. Check /api/rate-limit and /api/billing/spending-limit before making requests.

​
GET /api/ai/ghost/[id]/stream

Resume an existing AI conversation stream that was interrupted.
​
id
string
required
Context key for the conversation
​
resumeAt
number
Stream position to resume from (in bytes)

​
Response

​
stream
ReadableStream
Continuation of the existing stream from the specified position
const response = await fetch('/api/ai/ghost/pr-123/stream?resumeAt=2048', {
  method: 'GET'
});

const reader = response.body.getReader();
// Continue reading stream

​
POST /api/ai/command

Execute AI-powered GitHub commands via natural language (used in command palette).
​
messages
array
required
Array of conversation messages
​
pageContext
object
Current page context including:
​
page
string
Current page type (repo, issue, pr, etc.)
​
pathname
string
Current URL pathname
​
entities
array
Entities visible on current page
{
  type: string,
  id: string,
  name: string,
  [key: string]: unknown
}

​
Response

​
stream
ReadableStream
Streaming response with command execution results
const response = await fetch('/api/ai/command', {
  method: 'POST',
  headers: { 'Content-Type': 'application/json' },
  body: JSON.stringify({
    messages: [
      { role: 'user', content: 'Star this repository' }
    ],
    pageContext: {
      page: 'repo',
      pathname: '/repos/vercel/next.js',
      entities: [
        { type: 'repo', id: 'vercel/next.js', name: 'Next.js' }
      ]
    }
  })
});

​
POST /api/ai/commit-message

Generate AI-powered commit messages from file changes or PR information.
​
mode
string
Generation mode: "squash" for PR squash commits or omit for single file commits

​
For Single File Commits

​
filename
string
required
Name of the changed file
​
originalContent
string
required
Original file content
​
newContent
string
required
Modified file content

​
For PR Squash Commits

​
prTitle
string
required
Pull request title
​
prBody
string
Pull request description
​
prNumber
number
required
Pull request number
​
commits
array
Array of commit messages to squash (max 30)

​
Response

​
message
string
Generated commit message (for single file)
​
title
string
Generated commit title (for squash)
​
description
string
Generated commit description (for squash)
const response = await fetch('/api/ai/commit-message', {
  method: 'POST',
  headers: { 'Content-Type': 'application/json' },
  body: JSON.stringify({
    filename: 'auth.ts',
    originalContent: 'const auth = null;',
    newContent: 'const auth = new AuthService();'
  })
});

const data = await response.json();
// { message: "feat: initialize AuthService in auth.ts" }
This endpoint consumes AI credits. Maximum code length is determined by the model context window.

​
Chat History Endpoints

​
GET /api/ai/chat-history

Retrieve chat conversation history.
​
contextKey
string
Context key for a specific conversation
​
list
string
Set to "ghost" to list all ghost chat conversations

​
Response

​
conversation
object
Conversation metadata
​
id
string
Conversation ID
​
userId
string
User ID
​
chatType
string
Type of chat (ghost, command, etc.)
​
contextKey
string
Context identifier
​
title
string
Conversation title
​
activeStreamId
string
Active stream ID if streaming
​
createdAt
string
Creation timestamp
​
updatedAt
string
Last update timestamp
​
messages
array
Array of conversation messages
const response = await fetch('/api/ai/chat-history?contextKey=pr-123');
const data = await response.json();
/*
{
  conversation: {
    id: "conv_123",
    userId: "user_456",
    chatType: "ghost",
    contextKey: "pr-123",
    title: "Review PR #123",
    createdAt: "2024-01-15T10:00:00Z",
    updatedAt: "2024-01-15T10:05:00Z"
  },
  messages: [
    {
      id: "msg_1",
      role: "user",
      content: "Review this PR",
      createdAt: "2024-01-15T10:00:00Z"
    }
  ]
}
*/

​
POST /api/ai/chat-history

Save a message to a conversation.
​
contextKey
string
required
Context key for the conversation
​
chatType
string
required
Type of chat
​
message
object
required
Message to save

​
DELETE /api/ai/chat-history

Delete a conversation.
​
conversationId
string
required
ID of conversation to delete

​
Ghost Tab Management

​
GET /api/ai/ghost-tabs

Get the current state of Ghost chat tabs for the user.

​
Response

​
tabs
array
Array of active tabs
​
activeTabId
string
Currently active tab ID
​
counter
number
Tab counter for generating new tab IDs

​
POST /api/ai/ghost-tabs

Manage Ghost chat tabs.
​
action
string
required
Action to perform: "add", "close", "rename", or "switch"
​
tabId
string
Tab ID (required for all actions)
​
label
string
Tab label (required for add and rename)
​
counter
number
Counter value (required for add)
​
newDefault
string
New default tab ID (for close action)
const response = await fetch('/api/ai/ghost-tabs', {
  method: 'POST',
  headers: { 'Content-Type': 'application/json' },
  body: JSON.stringify({
    action: 'add',
    tabId: 'tab-1',
    label: 'New Chat',
    counter: 1
  })
});

Build docs developers (and LLMs) love

Get started for freeTalk to us