Screenpipe

https://docs.screenpipe.com
Admin
Screenpipe is a tool that records your screen and audio 24/7 locally, making your entire activity...
Tokens:122,230
Snippets:1,111
Trust Score:7.3
Update:2 months ago
Show doc for...
Tokens:
Raw
### Start local transcription server on macOS

Source: https://docs.screenpipe.com/meeting-transcription

Install and launch the speech server on Apple Silicon using Homebrew.

```bash
brew install speech
speech-server --port 8080
```

--------------------------------

### Example Continue Queries

Source: https://docs.screenpipe.com/continue

Example prompts to use within the Continue interface once the MCP server is configured.

```text
> @screenpipe what was I working on this morning?

> find the error I saw in my terminal and help me fix it

> what documentation was I reading about react hooks?
```

--------------------------------

### Setup OpenClaw with screenpipe

Source: https://docs.screenpipe.com/openclaw

Run this command to automatically install screenpipe skills and register the MCP server in OpenClaw's configuration.

```bash
npx -y screenpipe@latest agent setup openclaw
```

--------------------------------

### Workflow: Find Code Examples

Source: https://docs.screenpipe.com/msty

Example query to search screen history for specific code patterns.

```text
> search screenpipe for the python code I was looking at
> in the browser yesterday about asyncio patterns
```

--------------------------------

### Install and Enable a Pipe

Source: https://docs.screenpipe.com/openclaw

Use these commands to install and enable a pipe within the screenpipe application.

```bash
npx -y screenpipe@latest pipe install <path> && npx -y screenpipe@latest pipe enable <name>
```

```bash
bunx -y screenpipe@latest pipe install <path> && bunx -y screenpipe@latest pipe enable <name>
```

--------------------------------

### Start a Manual Meeting API Requests

Source: https://docs.screenpipe.com/api-reference/meetings/start-a-manual-meeting

Examples of how to trigger a manual meeting using various programming languages and tools.

```curl
curl --request POST \
  --url https://api.example.com/meetings/start \
  --header 'Content-Type: application/json' \
  --data '
{
  "app": "<string>",
  "title": "<string>",
  "attendees": "<string>"
}
'
```

```python
import requests

url = "https://api.example.com/meetings/start"

payload = {
    "app": "<string>",
    "title": "<string>",
    "attendees": "<string>"
}
headers = {"Content-Type": "application/json"}

response = requests.post(url, json=payload, headers=headers)

print(response.text)
```

```javascript
const options = {
  method: 'POST',
  headers: {'Content-Type': 'application/json'},
  body: JSON.stringify({app: '<string>', title: '<string>', attendees: '<string>'})
};

fetch('https://api.example.com/meetings/start', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));
```

```php
<?php

$curl = curl_init();

curl_setopt_array($curl, [
  CURLOPT_URL => "https://api.example.com/meetings/start",
  CURLOPT_RETURNTRANSFER => true,
  CURLOPT_ENCODING => "",
  CURLOPT_MAXREDIRS => 10,
  CURLOPT_TIMEOUT => 30,
  CURLOPT_HTTP_VERSION => CURL_HTTP_VERSION_1_1,
  CURLOPT_CUSTOMREQUEST => "POST",
  CURLOPT_POSTFIELDS => json_encode([
    'app' => '<string>',
    'title' => '<string>',
    'attendees' => '<string>'
  ]),
  CURLOPT_HTTPHEADER => [
    "Content-Type: application/json"
  ],
]);

$response = curl_exec($curl);
$err = curl_error($curl);

curl_close($curl);

if ($err) {
  echo "cURL Error #:" . $err;
} else {
  echo $response;
}
```

```go
package main

import (
	"fmt"
	"strings"
	"net/http"
	"io"
)

func main() {

	url := "https://api.example.com/meetings/start"

	payload := strings.NewReader("{\n  \"app\": \"<string>\",\n  \"title\": \"<string>\",\n  \"attendees\": \"<string>\"\n}")

	req, _ := http.NewRequest("POST", url, payload)

	req.Header.Add("Content-Type", "application/json")

	res, _ := http.DefaultClient.Do(req)

	defer res.Body.Close()
	body, _ := io.ReadAll(res.Body)

	fmt.Println(string(body))

}
```

```java
HttpResponse<String> response = Unirest.post("https://api.example.com/meetings/start")
  .header("Content-Type", "application/json")
  .body("{\n  \"app\": \"<string>\",\n  \"title\": \"<string>\",\n  \"attendees\": \"<string>\"\n}")
  .asString();
```

```ruby
require 'uri'
require 'net/http'

url = URI("https://api.example.com/meetings/start")

http = Net::HTTP.new(url.host, url.port)
http.use_ssl = true

request = Net::HTTP::Post.new(url)
request["Content-Type"] = 'application/json'
request.body = "{\n  \"app\": \"<string>\",\n  \"title\": \"<string>\",\n  \"attendees\": \"<string>\"\n}"

response = http.request(request)
puts response.read_body
```

--------------------------------

### POST /audio/start

Source: https://docs.screenpipe.com/cli-reference

Start audio recording.

```APIDOC
## POST /audio/start

### Description
Start audio recording.

### Method
POST

### Endpoint
/audio/start
```

--------------------------------

### Install Linux Dependencies

Source: https://docs.screenpipe.com/faq

Installs required system packages for Screenpipe on Ubuntu.

```bash
sudo apt install tesseract-ocr libxcb1
```

--------------------------------

### Install Linux Dependencies

Source: https://docs.screenpipe.com/getting-started

Installs required system libraries for audio and video processing on Linux distributions.

```bash
sudo apt install libasound2-dev ffmpeg  # ubuntu/debian
sudo dnf install alsa-lib ffmpeg         # fedora
```

--------------------------------

### Start local transcription server on Linux and Windows

Source: https://docs.screenpipe.com/meeting-transcription

Download models and start the transcription server on non-macOS platforms.

```bash
# Linux
speech download-models
speech serve
```

```powershell
# Windows PowerShell, from the extracted package's bin directory
.\speech_download_models.ps1
.\speech-server.exe
```

--------------------------------

### Start a Manual Meeting in Go

Source: https://docs.screenpipe.com/api-reference/meetings/start-a-manual-meeting

Uses the standard net/http library to send a POST request to the meeting start endpoint.

```go
package main

import (
	"fmt"
	"strings"
	"net/http"
	"io"
)

func main() {

	url := "https://api.example.com/meetings/start"

	payload := strings.NewReader("{\n  \"app\": \"<string>\",\n  \"title\": \"<string>\",\n  \"attendees\": \"<string>\"\n}")

	req, _ := http.NewRequest("POST", url, payload)

	req.Header.Add("Content-Type", "application/json")

	res, _ := http.DefaultClient.Do(req)

	defer res.Body.Close()
	body, _ := io.ReadAll(res.Body)

	fmt.Println(string(body))

}
```

--------------------------------

### Start Screenpipe Recording

Source: https://docs.screenpipe.com/copilot-cli

Initializes the Screenpipe recording process in the terminal.

```bash
npx -y screenpipe@latest record
```

--------------------------------

### POST /pipes/install

Source: https://docs.screenpipe.com/pipes

Installs a new pipe from a local directory path.

```APIDOC
## POST /pipes/install

### Description
Installs a pipe from the specified source directory.

### Method
POST

### Endpoint
http://localhost:3030/pipes/install

### Request Body
- **source** (string) - Required - Path to the pipe directory

### Request Example
curl -X POST http://localhost:3030/pipes/install -H "Authorization: Bearer $SCREENPIPE_LOCAL_API_KEY" -H "Content-Type: application/json" -d '{"source": "~/.screenpipe/pipes/my-pipe"}'
```

--------------------------------

### Cline Usage Examples

Source: https://docs.screenpipe.com/cline

Example prompts for querying screen history and terminal logs within the Cline interface.

```text
> I was reading documentation about async/await patterns earlier,
> find it and help me apply those patterns here

> what error messages have I seen in my terminal today?

> find the API response format I was looking at in the browser
```

--------------------------------

### POST /pipes/install

Source: https://docs.screenpipe.com/pipes

Install a new pipe from a remote URL.

```APIDOC
## POST /pipes/install

### Description
Install a pipe from a provided source URL.

### Method
POST

### Endpoint
http://localhost:3030/pipes/install

### Parameters
#### Request Body
- **source** (string) - Required - The URL pointing to the pipe definition file.
```

--------------------------------

### Start Audio Recording OpenAPI Specification

Source: https://docs.screenpipe.com/api-reference/audio/start-audio-recording

Defines the POST endpoint for starting audio recording within the Screenpipe API.

```yaml
openapi: 3.0.3
info:
  title: Screenpipe API
  version: 1.0.0
  description: >-
    Screenpipe captures everything you see, say, and hear on your computer. Use
    this API to search through captured content, manage recordings, and build
    AI-powered automations on top of your screen data.


    The server runs locally at `http://localhost:3030` by default.
servers: []
security: []
tags:
  - name: Search
    description: Search through captured screen and audio content
  - name: Frames
    description: Access captured screenshots and their extracted text
  - name: Elements
    description: Query captured UI accessibility tree data
  - name: Audio
    description: Manage audio recording devices
  - name: Vision
    description: Manage screen capture monitors
  - name: Meetings
    description: Detected and manual meeting transcriptions
  - name: Speakers
    description: Speaker identification and management
  - name: Memories
    description: AI-extracted knowledge from screen activity
  - name: Tags
    description: Tag content items for organization
  - name: Activity
    description: Activity summaries and analytics
  - name: Vault
    description: Encrypt/decrypt all data at rest
  - name: Cloud Sync
    description: Sync data across devices via cloud
  - name: Cloud Archive
    description: Archive old data to cloud storage
  - name: Data Retention
    description: Auto-delete old data locally
  - name: Data Management
    description: Manual data deletion and storage info
  - name: Database
    description: Direct database access
  - name: System
    description: Health checks and system status
  - name: Experimental
    description: Experimental/unstable endpoints
paths:
  /audio/start:
    post:
      tags:
        - Audio
      summary: Start audio recording
      operationId: routes_audio_start_audio
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema: {}
```

--------------------------------

### Configure Intune Install and Uninstall Commands

Source: https://docs.screenpipe.com/intune-deployment

Use these commands in the Intune Win32 app configuration to handle silent installation and uninstallation of the Screenpipe enterprise package.

```bash
powershell.exe -ExecutionPolicy Bypass -File install-screenpipe-enterprise.ps1
```

```bash
"%ProgramFiles%\screenpipe\uninstall.exe" /S
```

--------------------------------

### POST /audio/start

Source: https://docs.screenpipe.com/api-reference/audio/start-audio-recording

Starts the audio recording process on the local Screenpipe server.

```APIDOC
## POST /audio/start

### Description
Starts audio recording on the local Screenpipe instance.

### Method
POST

### Endpoint
/audio/start

### Response
#### Success Response (200)
- **OK** (object) - The recording has been successfully started.
```

--------------------------------

### GET /pipes

Source: https://docs.screenpipe.com/pipes

Retrieve a list of all installed pipes.

```APIDOC
## GET /pipes

### Description
List all installed pipes.

### Method
GET

### Endpoint
http://localhost:3030/pipes
```

--------------------------------

### Example Context-Aware Workflows

Source: https://docs.screenpipe.com/gemini-cli

Sample prompts demonstrating how to leverage screen history for coding tasks, debugging, and documentation recall.

```text
> I was looking at a react component earlier that handled
> form validation, find it and help me implement something similar
```

```text
> there was an error in my build output, find it and help me fix it
```

```text
> what did that API documentation say about rate limits?
```

--------------------------------

### Create and initialize a new pipe

Source: https://docs.screenpipe.com/pipes

Create a directory and a pipe.md configuration file, then install and test the pipe using the CLI.

```bash
mkdir -p ~/.screenpipe/pipes/my-pipe
cat > ~/.screenpipe/pipes/my-pipe/pipe.md << 'EOF'
---
schedule: every 30m
enabled: true
---

Summarize my screen activity for the last 30 minutes.
Query screenpipe at http://localhost:3030/search using the time range from the context header.
Authenticate with the SCREENPIPE_LOCAL_API_KEY environment variable.
Write the summary to ./output/<date>.md
EOF

# install + enable + test it from the CLI (no install needed — npx / bunx / bun x all work):
npx -y screenpipe@latest pipe install ~/.screenpipe/pipes/my-pipe
npx -y screenpipe@latest pipe enable my-pipe
npx -y screenpipe@latest pipe run my-pipe        # run once now to test
#
# (or use the desktop app: Pipes → My Pipes — or the authenticated REST API)
```

--------------------------------

### GET /frames/{id}/metadata

Source: https://docs.screenpipe.com/cli-reference

Get frame metadata.

```APIDOC
## GET /frames/{id}/metadata

### Description
Get frame metadata.

### Method
GET

### Endpoint
/frames/{id}/metadata

### Parameters
#### Path Parameters
- **id** (string) - Required - The frame identifier
```

--------------------------------

### Install and run Ollama model

Source: https://docs.screenpipe.com/ollama

Initializes the Ollama service and downloads the specified model for local execution.

```bash
# install from https://ollama.com then:
ollama run llama3.2
```

--------------------------------

### GET /frames/{id}/elements

Source: https://docs.screenpipe.com/cli-reference

Get UI elements for a frame.

```APIDOC
## GET /frames/{id}/elements

### Description
Get UI elements for a frame.

### Method
GET

### Endpoint
/frames/{id}/elements

### Parameters
#### Path Parameters
- **id** (string) - Required - The frame identifier
```

--------------------------------

### Example Workflows

Source: https://docs.screenpipe.com/continue

Specific use cases for leveraging screen context in coding tasks.

```text
> I was looking at how another project handles authentication,
> find that code and help me implement something similar
```

```text
> there was a typescript error earlier about types,
> find it and explain how to fix it
```

```text
> what did we discuss in the standup about the API changes?
> help me implement those changes
```

--------------------------------

### GET /frames/{id}/context

Source: https://docs.screenpipe.com/cli-reference

Get surrounding accessibility context for a frame.

```APIDOC
## GET /frames/{id}/context

### Description
Get surrounding accessibility context for a frame.

### Method
GET

### Endpoint
/frames/{id}/context

### Parameters
#### Path Parameters
- **id** (string) - Required - The frame identifier
```

--------------------------------

### POST /sync/init

Source: https://docs.screenpipe.com/api-reference/cloud-sync/initialize-cloud-sync

Initializes cloud synchronization at runtime with the required credentials and configuration.

```APIDOC
## POST /sync/init

### Description
Initialize sync at runtime with credentials.

### Method
POST

### Endpoint
/sync/init

### Request Body
- **token** (string) - Required - The authentication token.
- **password** (string) - Required - The user password.
- **machine_id** (string) - Required - The unique identifier for the machine.
- **sync_interval_secs** (integer) - Required - The interval in seconds for synchronization.

### Request Example
{
  "token": "string",
  "password": "string",
  "machine_id": "string",
  "sync_interval_secs": 0
}

### Response
#### Success Response (200)
- **success** (boolean) - Indicates if the initialization was successful.
- **is_new_user** (boolean) - Indicates if the user is new.
- **machine_id** (string) - The machine identifier associated with the sync.

#### Response Example
{
  "success": true,
  "is_new_user": true,
  "machine_id": "string"
}
```

--------------------------------

### POST /vault/setup

Source: https://docs.screenpipe.com/api-reference/vault/set-up-vault

Initializes the vault for data encryption at rest.

```APIDOC
## POST /vault/setup

### Description
Initializes the vault for data encryption at rest.

### Method
POST

### Endpoint
/vault/setup

### Request Body
- **password** (string) - Required - The password used to set up the vault.

### Request Example
{
  "password": "your-secure-password"
}

### Response
#### Success Response (200)
- **object** (object) - The response object confirming the vault setup.
```

--------------------------------

### List Installed Pipes

Source: https://docs.screenpipe.com/pipe-debugging

Confirms that a specific pipe is correctly installed and enabled in the system.

```bash
curl http://localhost:3030/pipes
```

--------------------------------

### OpenAPI Specification for Get Archive Status

Source: https://docs.screenpipe.com/api-reference/cloud-archive/get-archive-status

Defines the GET /archive/status endpoint and the ArchiveStatusResponse schema.

```yaml
openapi: 3.0.3
info:
  title: Screenpipe API
  version: 1.0.0
  description: >-
    Screenpipe captures everything you see, say, and hear on your computer. Use
    this API to search through captured content, manage recordings, and build
    AI-powered automations on top of your screen data.


    The server runs locally at `http://localhost:3030` by default.
servers: []
security: []
tags:
  - name: Search
    description: Search through captured screen and audio content
  - name: Frames
    description: Access captured screenshots and their extracted text
  - name: Elements
    description: Query captured UI accessibility tree data
  - name: Audio
    description: Manage audio recording devices
  - name: Vision
    description: Manage screen capture monitors
  - name: Meetings
    description: Detected and manual meeting transcriptions
  - name: Speakers
    description: Speaker identification and management
  - name: Memories
    description: AI-extracted knowledge from screen activity
  - name: Tags
    description: Tag content items for organization
  - name: Activity
    description: Activity summaries and analytics
  - name: Vault
    description: Encrypt/decrypt all data at rest
  - name: Cloud Sync
    description: Sync data across devices via cloud
  - name: Cloud Archive
    description: Archive old data to cloud storage
  - name: Data Retention
    description: Auto-delete old data locally
  - name: Data Management
    description: Manual data deletion and storage info
  - name: Database
    description: Direct database access
  - name: System
    description: Health checks and system status
  - name: Experimental
    description: Experimental/unstable endpoints
paths:
  /archive/status:
    get:
      tags:
        - Cloud Archive
      summary: Get archive status
      description: GET /archive/status — return current state.
      operationId: archive_archive_status
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ArchiveStatusResponse'
components:
  schemas:
    ArchiveStatusResponse:
      type: object
      properties:
        enabled:
          type: boolean
        retention_days:
          type: integer
        watermark:
          nullable: true
          type: string
        last_cleanup:
          nullable: true
          type: string
        last_error:
          nullable: true
          type: string
        storage_used:
          type: integer
        storage_limit:
          type: integer
        pending_count:
          type: integer
        is_uploading:
          type: boolean
        chunks_uploaded:
          type: integer
        media_files_uploaded:
          type: integer
        media_files_pending:
          type: integer
      required:
        - enabled
        - retention_days
        - watermark
        - last_cleanup
        - last_error
        - storage_used
        - storage_limit
        - pending_count
        - is_uploading
        - chunks_uploaded
        - media_files_uploaded
        - media_files_pending
```

--------------------------------

### Start Screenpipe with LAN access

Source: https://docs.screenpipe.com/search-screen-history

Use this command to bind the server to 0.0.0.0, allowing access from other devices on the local network.

```bash
npx -y screenpipe@latest --listen-on-lan
```

--------------------------------

### POST /audio/device/start

Source: https://docs.screenpipe.com/api-reference/audio/start-recording-a-specific-audio-device

Starts recording from a specific audio device identified by its name.

```APIDOC
## POST /audio/device/start

### Description
Starts recording a specific audio device.

### Method
POST

### Endpoint
/audio/device/start

### Request Body
- **device_name** (string) - Required - The name of the audio device to start recording.

### Request Example
{
  "device_name": "string"
}

### Response
#### Success Response (200)
- **success** (boolean) - Indicates if the operation was successful.
- **message** (string) - A message describing the result of the operation.

#### Response Example
{
  "success": true,
  "message": "string"
}
```

--------------------------------

### Install pi-agent via CLI

Source: https://docs.screenpipe.com/faq

Use this command to manually install or update the pi-agent runtime when it is missing or failing.

```bash
npx -y screenpipe@latest pipe install pi-agent
```

--------------------------------

### Install and Enable Second Brain Pipe

Source: https://docs.screenpipe.com/second-brain

Commands to install and enable the second-brain pipe within the Screenpipe environment.

```bash
npx -y screenpipe@latest pipe install ~/.screenpipe/pipes/second-brain && npx -y screenpipe@latest pipe enable second-brain
```

--------------------------------

### Msty Usage Examples

Source: https://docs.screenpipe.com/msty

Example natural language queries for Msty to search screen history and terminal activity.

```text
> what was I working on this morning?

> find that documentation about async/await patterns I was reading earlier

> what error messages have I seen in my terminal today?

> search for the API response format I was looking at in the browser
```

--------------------------------

### Compare Prompt Quality

Source: https://docs.screenpipe.com/pipe-debugging

Examples of weak versus strong prompt structures for better output control.

```markdown
Tell me what I did today.
```

```markdown
Search the last 24 hours of screenpipe data.
Write a markdown summary to ~/Documents/daily-screenpipe-summary.md.
Include:
- accomplishments
- meetings
- open loops
- timeline links for the 5 most important moments
If there is no data, say which API call returned empty.
```

--------------------------------

### Setup other agents with screenpipe

Source: https://docs.screenpipe.com/openclaw

Use the same command structure to wire up other supported agents.

```bash
npx -y screenpipe@latest agent setup <hermes | claude-code | claude-desktop | codex | cursor | windsurf>
```

--------------------------------

### OpenAPI Specification for Get Frame Context

Source: https://docs.screenpipe.com/api-reference/frames/get-frame-context

Defines the GET endpoint for retrieving frame context, including the request parameter and the response schema.

```yaml
openapi: 3.0.3
info:
  title: Screenpipe API
  version: 1.0.0
  description: >-
    Screenpipe captures everything you see, say, and hear on your computer. Use
    this API to search through captured content, manage recordings, and build
    AI-powered automations on top of your screen data.


    The server runs locally at `http://localhost:3030` by default.
servers: []
security: []
tags:
  - name: Search
    description: Search through captured screen and audio content
  - name: Frames
    description: Access captured screenshots and their extracted text
  - name: Elements
    description: Query captured UI accessibility tree data
  - name: Audio
    description: Manage audio recording devices
  - name: Vision
    description: Manage screen capture monitors
  - name: Meetings
    description: Detected and manual meeting transcriptions
  - name: Speakers
    description: Speaker identification and management
  - name: Memories
    description: AI-extracted knowledge from screen activity
  - name: Tags
    description: Tag content items for organization
  - name: Activity
    description: Activity summaries and analytics
  - name: Vault
    description: Encrypt/decrypt all data at rest
  - name: Cloud Sync
    description: Sync data across devices via cloud
  - name: Cloud Archive
    description: Archive old data to cloud storage
  - name: Data Retention
    description: Auto-delete old data locally
  - name: Data Management
    description: Manual data deletion and storage info
  - name: Database
    description: Direct database access
  - name: System
    description: Health checks and system status
  - name: Experimental
    description: Experimental/unstable endpoints
paths:
  /frames/{frame_id}/context:
    get:
      tags:
        - Frames
      summary: Get frame context
      description: |-
        Get frame context: accessibility text, tree nodes, and extracted URLs.
        Falls back to OCR data for legacy frames without accessibility data.
      operationId: routes_frames_get_frame_context
      parameters:
        - name: frame_id
          schema:
            type: integer
          in: path
          required: true
          style: simple
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/FrameContextResponse'
components:
  schemas:
    FrameContextResponse:
      description: >-
        Response type for frame context endpoint (accessibility-first, OCR
        fallback)
      type: object
      properties:
        frame_id:
          type: integer
        text:
          nullable: true
          type: string
        nodes:
          type: array
          items:
            $ref: '#/components/schemas/AccessibilityNode'
        urls:
          type: array
          items:
            type: string
        text_source:
          type: string
      required:
        - frame_id
        - text
        - nodes
        - urls
        - text_source
    AccessibilityNode:
      description: A node from the accessibility tree
      type: object
      properties:
        role:
          type: string
        text:
          type: string
        depth:
          type: integer
        bounds:
          $ref: '#/components/schemas/AccessibilityNodeBounds'
        properties:
          nullable: true
          type: object
      required:
        - role
        - text
        - depth
    AccessibilityNodeBounds:
      description: Bounding box for an accessibility node (0-1 normalized to monitor)
      type: object
      properties:
        left:
          type: number
        top:
          type: number
        width:
          type: number
        height:
          type: number
      required:
        - left
        - top
        - width
        - height
```

--------------------------------

### POST /add

Source: https://docs.screenpipe.com/api-reference/database/add-content-to-database

Manually insert screen or audio content into the database.

```APIDOC
## POST /add

### Description
Manually insert screen or audio content into the database.

### Method
POST

### Endpoint
/add

### Request Body
- **device_name** (string) - Required - The name of the device.
- **content** (object) - Required - The content data to be added.

### Response
#### Success Response (200)
- **success** (boolean) - Indicates if the operation was successful.
- **message** (string) - A message regarding the operation status.
```

--------------------------------

### Start Screenpipe HTTP MCP Server

Source: https://docs.screenpipe.com/msty

Run this command in your terminal to start the Screenpipe MCP server in HTTP mode on port 3031.

```bash
npx -y screenpipe-mcp-http --port 3031
```

--------------------------------

### OpenAPI definition for Get frame metadata

Source: https://docs.screenpipe.com/api-reference/frames/get-frame-metadata

The OpenAPI specification for the GET /frames/{frame_id}/metadata endpoint, including the response schema for frame metadata.

```yaml
openapi: 3.0.3
info:
  title: Screenpipe API
  version: 1.0.0
  description: >-
    Screenpipe captures everything you see, say, and hear on your computer. Use
    this API to search through captured content, manage recordings, and build
    AI-powered automations on top of your screen data.


    The server runs locally at `http://localhost:3030` by default.
servers: []
security: []
tags:
  - name: Search
    description: Search through captured screen and audio content
  - name: Frames
    description: Access captured screenshots and their extracted text
  - name: Elements
    description: Query captured UI accessibility tree data
  - name: Audio
    description: Manage audio recording devices
  - name: Vision
    description: Manage screen capture monitors
  - name: Meetings
    description: Detected and manual meeting transcriptions
  - name: Speakers
    description: Speaker identification and management
  - name: Memories
    description: AI-extracted knowledge from screen activity
  - name: Tags
    description: Tag content items for organization
  - name: Activity
    description: Activity summaries and analytics
  - name: Vault
    description: Encrypt/decrypt all data at rest
  - name: Cloud Sync
    description: Sync data across devices via cloud
  - name: Cloud Archive
    description: Archive old data to cloud storage
  - name: Data Retention
    description: Auto-delete old data locally
  - name: Data Management
    description: Manual data deletion and storage info
  - name: Database
    description: Direct database access
  - name: System
    description: Health checks and system status
  - name: Experimental
    description: Experimental/unstable endpoints
paths:
  /frames/{frame_id}/metadata:
    get:
      tags:
        - Frames
      summary: Get frame metadata
      description: >-
        Get frame metadata (timestamp) for deep link navigation.
        screenpipe://frame/123 → resolve to timestamp.
      operationId: routes_frames_get_frame_metadata
      parameters:
        - name: frame_id
          schema:
            type: integer
          in: path
          required: true
          style: simple
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/FrameMetadataResponse'
components:
  schemas:
    FrameMetadataResponse:
      description: Response type for frame metadata (used by deep link navigation)
      type: object
      properties:
        frame_id:
          type: integer
        timestamp:
          type: string
          format: date-time
      required:
        - frame_id
        - timestamp
```

--------------------------------

### Build Screenpipe MCP from source

Source: https://docs.screenpipe.com/mcp-server

Clone the repository and build the package to use a local version of the MCP server.

```bash
git clone https://github.com/screenpipe/screenpipe
cd screenpipe/packages/screenpipe-mcp
npm install && npm run build
```

--------------------------------

### OpenAPI definition for get frame text and bounds

Source: https://docs.screenpipe.com/api-reference/frames/get-frame-text-and-bounds

Defines the GET /frames/{frame_id}/text endpoint and the associated response schemas for text positions and bounds.

```yaml
openapi: 3.0.3
info:
  title: Screenpipe API
  version: 1.0.0
  description: >-
    Screenpipe captures everything you see, say, and hear on your computer. Use
    this API to search through captured content, manage recordings, and build
    AI-powered automations on top of your screen data.


    The server runs locally at `http://localhost:3030` by default.
servers: []
security: []
tags:
  - name: Search
    description: Search through captured screen and audio content
  - name: Frames
    description: Access captured screenshots and their extracted text
  - name: Elements
    description: Query captured UI accessibility tree data
  - name: Audio
    description: Manage audio recording devices
  - name: Vision
    description: Manage screen capture monitors
  - name: Meetings
    description: Detected and manual meeting transcriptions
  - name: Speakers
    description: Speaker identification and management
  - name: Memories
    description: AI-extracted knowledge from screen activity
  - name: Tags
    description: Tag content items for organization
  - name: Activity
    description: Activity summaries and analytics
  - name: Vault
    description: Encrypt/decrypt all data at rest
  - name: Cloud Sync
    description: Sync data across devices via cloud
  - name: Cloud Archive
    description: Archive old data to cloud storage
  - name: Data Retention
    description: Auto-delete old data locally
  - name: Data Management
    description: Manual data deletion and storage info
  - name: Database
    description: Direct database access
  - name: System
    description: Health checks and system status
  - name: Experimental
    description: Experimental/unstable endpoints
paths:
  /frames/{frame_id}/text:
    get:
      tags:
        - Frames
      summary: Get frame text and bounds
      description: >-
        Get frame text positions with bounding boxes for a specific frame.

        Uses accessibility tree node bounds when available, and OCR fallback
        positions for visual-only or legacy frames.

        Both OCR and accessibility bounds are normalized to 0-1 relative to the

        monitor (full-screen capture), so they align correctly with the
        screenshot.
      operationId: routes_frames_get_frame_text_data
      parameters:
        - name: frame_id
          schema:
            type: integer
          in: path
          required: true
          style: simple
        - name: query
          schema:
            nullable: true
            type: string
          in: query
          style: form
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/FrameTextResponse'
components:
  schemas:
    FrameTextResponse:
      description: Response type for frame OCR data endpoint
      type: object
      properties:
        frame_id:
          type: integer
        text_positions:
          type: array
          items:
            $ref: '#/components/schemas/TextPosition'
      required:
        - frame_id
        - text_positions
    TextPosition:
      type: object
      properties:
        text:
          type: string
        confidence:
          type: number
        bounds:
          $ref: '#/components/schemas/TextBounds'
      required:
        - text
        - confidence
        - bounds
    TextBounds:
      type: object
      properties:
        left:
          type: number
        top:
          type: number
        width:
          type: number
        height:
          type: number
      required:
        - left
        - top
        - width
        - height
```

--------------------------------

### Start a Manual Meeting in Ruby

Source: https://docs.screenpipe.com/api-reference/meetings/start-a-manual-meeting

Uses the Net::HTTP library to initiate a POST request with SSL enabled.

```ruby
require 'uri'
require 'net/http'

url = URI("https://api.example.com/meetings/start")

http = Net::HTTP.new(url.host, url.port)
http.use_ssl = true

request = Net::HTTP::Post.new(url)
request["Content-Type"] = 'application/json'
request.body = "{\n  \"app\": \"<string>\",\n  \"title\": \"<string>\",\n  \"attendees\": \"<string>\"\n}"

response = http.request(request)
puts response.read_body
```

--------------------------------

### OpenAPI Specification for Get Frame by ID

Source: https://docs.screenpipe.com/api-reference/frames/get-frame-by-id

Defines the GET /frames/{frame_id} endpoint within the Screenpipe API, specifying the required frame_id parameter and the operation ID.

```yaml
openapi: 3.0.3
info:
  title: Screenpipe API
  version: 1.0.0
  description: >-
    Screenpipe captures everything you see, say, and hear on your computer. Use
    this API to search through captured content, manage recordings, and build
    AI-powered automations on top of your screen data.


    The server runs locally at `http://localhost:3030` by default.
servers: []
security: []
tags:
  - name: Search
    description: Search through captured screen and audio content
  - name: Frames
    description: Access captured screenshots and their extracted text
  - name: Elements
    description: Query captured UI accessibility tree data
  - name: Audio
    description: Manage audio recording devices
  - name: Vision
    description: Manage screen capture monitors
  - name: Meetings
    description: Detected and manual meeting transcriptions
  - name: Speakers
    description: Speaker identification and management
  - name: Memories
    description: AI-extracted knowledge from screen activity
  - name: Tags
    description: Tag content items for organization
  - name: Activity
    description: Activity summaries and analytics
  - name: Vault
    description: Encrypt/decrypt all data at rest
  - name: Cloud Sync
    description: Sync data across devices via cloud
  - name: Cloud Archive
    description: Archive old data to cloud storage
  - name: Data Retention
    description: Auto-delete old data locally
  - name: Data Management
    description: Manual data deletion and storage info
  - name: Database
    description: Direct database access
  - name: System
    description: Health checks and system status
  - name: Experimental
    description: Experimental/unstable endpoints
paths:
  /frames/{frame_id}:
    get:
      tags:
        - Frames
      summary: Get frame by ID
      description: Returns a captured screenshot frame with optional base64 image data.
      operationId: routes_frames_get_frame_data
      parameters:
        - name: frame_id
          schema:
            type: integer
          in: path
          required: true
          style: simple
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema: {}
```

--------------------------------

### Example Workflow Queries

Source: https://docs.screenpipe.com/claude-code

Natural language prompts for Claude Code to interact with Screenpipe data.

```text
> I was reading a blog post about rust async earlier today,
> search screenpipe and summarize the key points
```

```text
> search my audio transcriptions for what was discussed in standup
> about the API refactor, then help me implement it
```

```text
> I saw an error message flash on screen, search screenpipe
> to find it and help me fix the issue
```

```text
> search screenpipe for the python code I was looking at
> in the browser yesterday about asyncio patterns
```

```text
> what did I write in Notion this morning? search for my
> keyboard input using content_type=input
```

```text
> I copied something important earlier, search clipboard
> events to find it
```

--------------------------------

### GET /meetings

Source: https://docs.screenpipe.com/cli-reference

List all meetings.

```APIDOC
## GET /meetings

### Description
List all meetings.

### Method
GET

### Endpoint
/meetings
```

--------------------------------

### GET /search/keyword

Source: https://docs.screenpipe.com/cli-reference

Perform a keyword search.

```APIDOC
## GET /search/keyword

### Description
Perform a keyword search.

### Method
GET

### Endpoint
/search/keyword
```

--------------------------------

### Query Gemini CLI with Screen Context

Source: https://docs.screenpipe.com/gemini-cli

Example prompts for interacting with Gemini CLI after configuring the screenpipe MCP server.

```bash
gemini

> what was I working on this morning?

> find the error message I saw in my terminal earlier

> summarize the documentation I was reading about kubernetes
```

--------------------------------

### GET /health

Source: https://docs.screenpipe.com/cli-reference

Perform a server health check.

```APIDOC
## GET /health

### Description
Perform a server health check.

### Method
GET

### Endpoint
/health
```

--------------------------------

### OpenAPI definition for starting an audio device

Source: https://docs.screenpipe.com/api-reference/audio/start-recording-a-specific-audio-device

Defines the POST request structure for the /audio/device/start endpoint, requiring a device_name in the request body.

```yaml
openapi: 3.0.3
info:
  title: Screenpipe API
  version: 1.0.0
  description: >-
    Screenpipe captures everything you see, say, and hear on your computer. Use
    this API to search through captured content, manage recordings, and build
    AI-powered automations on top of your screen data.


    The server runs locally at `http://localhost:3030` by default.
servers: []
security: []
tags:
  - name: Search
    description: Search through captured screen and audio content
  - name: Frames
    description: Access captured screenshots and their extracted text
  - name: Elements
    description: Query captured UI accessibility tree data
  - name: Audio
    description: Manage audio recording devices
  - name: Vision
    description: Manage screen capture monitors
  - name: Meetings
    description: Detected and manual meeting transcriptions
  - name: Speakers
    description: Speaker identification and management
  - name: Memories
    description: AI-extracted knowledge from screen activity
  - name: Tags
    description: Tag content items for organization
  - name: Activity
    description: Activity summaries and analytics
  - name: Vault
    description: Encrypt/decrypt all data at rest
  - name: Cloud Sync
    description: Sync data across devices via cloud
  - name: Cloud Archive
    description: Archive old data to cloud storage
  - name: Data Retention
    description: Auto-delete old data locally
  - name: Data Management
    description: Manual data deletion and storage info
  - name: Database
    description: Direct database access
  - name: System
    description: Health checks and system status
  - name: Experimental
    description: Experimental/unstable endpoints
paths:
  /audio/device/start:
    post:
      tags:
        - Audio
      summary: Start recording a specific audio device
      operationId: routes_audio_start_audio_device
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/AudioDeviceControlRequest'
        required: true
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/AudioDeviceControlResponse'
components:
  schemas:
    AudioDeviceControlRequest:
      type: object
      properties:
        device_name:
          type: string
      required:
        - device_name
    AudioDeviceControlResponse:
      type: object
      properties:
        success:
          type: boolean
        message:
          type: string
      required:
        - success
        - message
```

--------------------------------

### Configure Screenpipe MCP in Continue

Source: https://docs.screenpipe.com/continue

Add the Screenpipe MCP server to your Continue configuration file to enable screen context access.

```json
{
  "mcpServers": [
    {
      "name": "screenpipe",
      "command": "npx",
      "args": ["-y", "screenpipe-mcp"]
    }
  ]
}
```

--------------------------------

### GET /elements

Source: https://docs.screenpipe.com/cli-reference

Search structured UI elements.

```APIDOC
## GET /elements

### Description
Search structured UI elements.

### Method
GET

### Endpoint
/elements
```
Tokens:
Raw