Hindsight

https://hindsight.vectorize.io
Admin
Hindsight is a state-of-the-art memory system for AI agents that enables temporal reasoning, fact...
Tokens:1,551,335
Snippets:11,877
Trust Score:9.4
Update:9 hours ago
Show doc for...
Tokens:
Raw
### Install Hindsight plugin and run memory setup

Source: https://hindsight.vectorize.io/blog/2026/09/25/hindsight-hermes-plugin-catalog

Run these two commands to install the Hindsight plugin from the catalog and start the memory setup wizard. The wizard installs dependencies via uv, walks through configuration, and offers to seed the bank with a starter template. It warns before overwriting an existing bank. Do not skip the second command.

```bash
hermes plugins install hindsight
hermes memory setup    # select "hindsight"
```

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

### Install and run setup wizard

Source: https://hindsight.vectorize.io/blog/2026/05/27/openclaw-codebase-memory

Install the @vectorize-io/hindsight-openclaw plugin and run the setup wizard. The wizard offers three install modes: Cloud, External API, and Embedded daemon.

```bash
openclaw plugins install @vectorize-io/hindsight-openclaw
npx --package @vectorize-io/hindsight-openclaw hindsight-openclaw-setup
```

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

### Start Demo Frontend

Source: https://hindsight.vectorize.io/cookbook/applications/deliveryman-demo

Installs npm dependencies and starts the frontend development server on http://localhost:5173.

```bash
cd hindsight-cookbook/deliveryman-demo/frontend
npm install
npm run dev
# Runs on http://localhost:5173
```

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

### Show hindsight-nemoclaw setup help

Source: https://hindsight.vectorize.io/blog/sandboxed-agent-persistent-memory-nemoclaw

Displays usage information for the setup command. Run this to get started with the setup process.

```bash
npx @vectorize-io/hindsight-nemoclaw setup --help
```

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

### Start TimescaleDB + DiskANN docker-compose example

Source: https://hindsight.vectorize.io/blog/2026/02/18/version-0-4-12

Use this command to start the TimescaleDB + DiskANN docker-compose example for self-hosted deployments.

```bash
# Use the TimescaleDB + DiskANN docker-compose example
docker compose -f docker/docker-compose/timescale/docker-compose.yml up
```

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

### Install Dependencies and Run Dev Server

Source: https://hindsight.vectorize.io/cookbook/applications/chat-sdk-multi-platform

Installs npm packages and starts the development server.

```bash
npm install
npm run dev
```

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

### Backend setup commands

Source: https://hindsight.vectorize.io/cookbook/applications/cable-co

Commands to set up and run the backend server on port 8002. Includes creating a virtual environment, installing dependencies, creating a .env file with API credentials, and starting the backend.

```bash
cd backend

# Create and activate a virtual environment
python -m venv venv
source venv/bin/activate

# Install dependencies
pip install -r requirements.txt

# Create a .env file with your credentials
cat > .env << 'EOF'
OPENAI_API_KEY=sk-your-openai-key
HINDSIGHT_API_KEY=hsk_your-hindsight-key
HINDSIGHT_API_URL=https://api.hindsight.vectorize.io
HINDSIGHT_BANK_NAME=cable-connect-demo
EOF

# Start the backend (port 8002)
./run.sh
```

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

### Install and Initialize Hindsight Zed Globally

Source: https://hindsight.vectorize.io/sdks/integrations/zed

Install the CLI globally with npm, then run the init command. This provides a persistent command for setup.

```bash
npm install -g @vectorize-io/hindsight-zed
hindsight-zed init --api-token YOUR_HINDSIGHT_API_KEY --bank-id my-memory
```

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

### Manual setup commands

Source: https://hindsight.vectorize.io/cookbook/applications/hindsight-litellm-demo

Installs required Python packages and runs the Streamlit app. Use this if you prefer manual setup over run.sh.

```bash
# Install dependencies
pip install streamlit litellm

# Install Hindsight packages
pip install hindsight-client hindsight-litellm

# Run the app
streamlit run app.py
```

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

### Install Hindsight OpenClaw plugin and start gateway

Source: https://hindsight.vectorize.io/guides/2026/04/15/guide-openclaw-per-user-memory-across-channels-setup

Run these commands to install the @vectorize-io/hindsight-openclaw plugin, run the setup script, and start the OpenClaw gateway. Requires OpenClaw installed and running, and a working Hindsight backend.

```bash
openclaw plugins install @vectorize-io/hindsight-openclaw
npx --package @vectorize-io/hindsight-openclaw hindsight-openclaw-setup
openclaw gateway
```

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

### Configure environment

Source: https://hindsight.vectorize.io/cookbook/applications/sanity-blog-memory

Copies the example configuration file and opens it for editing.

```bash
# Copy example config
cp .env.example .env

# Edit with your values
nano .env
```

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

### Self-hosted Hindsight server setup and install

Source: https://hindsight.vectorize.io/blog/2026/07/23/roo-code-persistent-memory

For self-hosted deployments, start the Hindsight API server first, then point the installer at it with --api-url. Requires pip install hindsight-all and setting HINDSIGHT_API_LLM_API_KEY.

```bash
pip install hindsight-all
export HINDSIGHT_API_LLM_API_KEY=your-openai-key
hindsight-api                                   # serves http://localhost:8888

hindsight-roo-code install --api-url http://localhost:8888
```

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

### Non-interactive setup wizard examples

Source: https://hindsight.vectorize.io/sdks/integrations/openclaw

Shows various non-interactive setup commands for CI and scripted environments, covering Cloud, External API, and Embedded daemon modes with inline or env-var credentials.

```bash
# Cloud — inline token (simplest)
npx --package @vectorize-io/hindsight-openclaw hindsight-openclaw-setup \
    --mode cloud --token hsk_your_cloud_token

# Cloud — SecretRef (read from env at gateway startup)
npx --package @vectorize-io/hindsight-openclaw hindsight-openclaw-setup \
    --mode cloud --token-env HINDSIGHT_CLOUD_TOKEN

# External API (no auth)
npx --package @vectorize-io/hindsight-openclaw hindsight-openclaw-setup \
    --mode api --api-url https://mcp.hindsight.example.com --no-token

# Embedded daemon with OpenAI — inline API key
npx --package @vectorize-io/hindsight-openclaw hindsight-openclaw-setup \
    --mode embedded --provider openai --api-key sk-...

# Embedded daemon with OpenAI — SecretRef
npx --package @vectorize-io/hindsight-openclaw hindsight-openclaw-setup \
    --mode embedded --provider openai --api-key-env OPENAI_API_KEY

# Embedded daemon with Claude Code (no API key needed)
npx --package @vectorize-io/hindsight-openclaw hindsight-openclaw-setup \
    --mode embedded --provider claude-code
```

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

### Start OpenCode

Source: https://hindsight.vectorize.io/blog/2026/04/20/opencode-persistent-memory

Launch OpenCode to activate the plugin and begin capturing context.

```bash
opencode
```

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

### Self-hosted Hindsight setup commands

Source: https://hindsight.vectorize.io/blog/2026/05/26/paperclip-persistent-memory

Commands to install Hindsight, set the OpenAI API key, and start the API server for self-hosting.

```bash
pip install hindsight-all
export HINDSIGHT_API_LLM_API_KEY=your-openai-key
hindsight-api
```

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

### Self-hosted Hindsight API setup

Source: https://hindsight.vectorize.io/blog/2026/06/09/cline-persistent-memory

Install and start the self-hosted API locally. Requires setting HINDSIGHT_API_LLM_API_KEY and running hindsight-api on http://localhost:8888.

```bash
pip install hindsight-all
export HINDSIGHT_API_LLM_API_KEY=your-openai-key
hindsight-api  # http://localhost:8888
```

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

### Set Up Demo Backend Virtual Environment

Source: https://hindsight.vectorize.io/cookbook/applications/deliveryman-demo

Creates a Python virtual environment and installs dependencies for the demo backend.

```bash
cd hindsight-cookbook/deliveryman-demo/backend

# Create virtual environment and install dependencies
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
```

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

### Frontend setup commands

Source: https://hindsight.vectorize.io/cookbook/applications/cable-co

Commands to set up and run the frontend dev server on port 5173. Includes installing dependencies and starting the dev server.

```bash
cd frontend

# Install dependencies
npm install

# Start the dev server (port 5173)
npm run dev
```

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

### Custom observation scopes for multi-tenant

Source: https://hindsight.vectorize.io/best-practices

Recommended for multi-tenant setups to control which tag combinations get their own observation pass. This example scopes observations to user-level, team-level, and combined.

```python
# Observations scoped to: user-level, team-level, and combined
observation_scopes=[
    ["user:alice"],
    ["team:engineering"],
    ["user:alice", "team:engineering"],
]
```

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

### Launch control center and configure with UI

Source: https://hindsight.vectorize.io/sdks/embed

Start the browser-based control center or use the browser wizard during setup. The control center listens on localhost only and prints a tokenized URL.

```bash
# Launch the control center and open the browser
hindsight-embed control start

# Or use the browser wizard instead of terminal prompts during setup
hindsight-embed configure --ui
```

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

### start()

Source: https://hindsight.vectorize.io/sdks/hindsight-all-npm

Configures the profile, spawns the daemon, and waits for the /health endpoint. It is idempotent and safe to re-run.

```APIDOC
## start()

### Description
Configure profile, spawn the daemon, wait for `/health`. Idempotent — safe to re-run.

### Method
start()

### Returns
`Promise<void>`

### Parameters
None

### Response
Resolves when the daemon is healthy.
```

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

### Get Knowledge Base Tree - Response Example

Source: https://hindsight.vectorize.io/developer/api/knowledge-pages

Example JSON response from the Get Knowledge Base Tree endpoint, showing a folder with a page child.

```json
{
  "roots": [
    {
      "id": "kf-9f2c...",
      "kind": "folder",
      "name": "Operations",
      "parent_id": null,
      "mental_model_id": null,
      "managed": false,
      "description": null,
      "tags": [],
      "timestamp": "2026-08-01T11:04:02+00:00",
      "is_stale": null,
      "children": [
        {
          "id": "kp-2e85...",
          "kind": "page",
          "name": "Deploying the API",
          "parent_id": "kf-9f2c...",
          "mental_model_id": "mm-77ab...",
          "managed": false,
          "description": "How is the API deployed?",
          "tags": ["ops"],
          "timestamp": "2026-08-03T09:12:44+00:00",
          "is_stale": true,
          "children": []
        }
      ]
    }
  ]
}
```

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

### Demo Backend Environment Configuration

Source: https://hindsight.vectorize.io/cookbook/applications/deliveryman-demo

Example backend/.env configuration with API keys and Hindsight API URL.

```bash
OPENAI_API_KEY=<your-openai-api-key>
GROQ_API_KEY=<your-groq-api-key>
HINDSIGHT_API_URL=http://localhost:8888
LLM_MODEL=openai/gpt-4o
```

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

### Run the App with npm

Source: https://hindsight.vectorize.io/cookbook/applications/chat-memory-cloud

Start the development server. After running, open http://localhost:3000 in your browser.

```bash
npm run dev
```

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

### Install and start Hindsight API server

Source: https://hindsight.vectorize.io/blog/2026/07/21/inkling-agent-memory-hindsight

Install the hindsight-api package and start the server, which serves on http://localhost:8888. The database is an embedded Postgres, so only extraction and reflection calls leave your machine.

```bash
pip install hindsight-api
hindsight-api        # serves http://localhost:8888
```
Tokens:
Raw