diff --git a/.readthedocs.yaml b/.readthedocs.yaml
new file mode 100644
index 0000000..520ddfe
--- /dev/null
+++ b/.readthedocs.yaml
@@ -0,0 +1,9 @@
+version: 2
+
+build:
+  os: ubuntu-22.04
+  tools:
+    python: "3.12"
+
+mkdocs:
+  configuration: mkdocs.yml
diff --git a/docs/cli.md b/docs/cli.md
new file mode 100644
index 0000000..234c056
--- /dev/null
+++ b/docs/cli.md
@@ -0,0 +1,72 @@
+# CLI Reference
+
+## loadforge run
+
+Run a load test from a scenario file.
+
+```bash
+loadforge run [scenario.yaml] [flags]
+```
+
+| Flag | Short | Description |
+|------|-------|-------------|
+| `--workers` | `-w` | Override worker count from config |
+| `--duration` | `-d` | Override duration (e.g. `30s`, `2m`) |
+| `--output` | `-o` | Save report to a JSON file |
+| `--var` | | Set a variable (`key=value`) |
+| `--env-file` | | Load variables from a `.env` file |
+| `--no-ui` | | Plain text output instead of the terminal UI |
+| `--verbose` | `-v` | Enable verbose logging |
+
+### Examples
+
+```bash
+# Run with defaults from the file
+loadforge run scenario.yaml
+
+# Override workers and duration
+loadforge run scenario.yaml --workers 50 --duration 2m
+
+# Inject a variable
+loadforge run scenario.yaml --var BASE_URL=https://staging.example.com
+
+# Load variables from .env
+loadforge run scenario.yaml --env-file .env
+```
+
+---
+
+## loadforge validate
+
+Validate a scenario file without running it.
+
+```bash
+loadforge validate [scenario.yaml]
+```
+
+---
+
+## loadforge version
+
+Print the installed version.
+
+```bash
+loadforge version
+```
+
+---
+
+## loadforge --uninstall
+
+Stop the web service and remove all LoadForge files from the system.
+
+```bash
+sudo loadforge --uninstall
+```
+
+This removes:
+
+- `/usr/local/bin/loadforge`
+- `/usr/local/bin/loadforge-web`
+- `~/.loadforge/` (config, history, logs)
+- The background service (systemd or launchd)
diff --git a/docs/configuration/assertions.md b/docs/configuration/assertions.md
new file mode 100644
index 0000000..2c3ffcf
--- /dev/null
+++ b/docs/configuration/assertions.md
@@ -0,0 +1,66 @@
+# Assertions
+
+Assertions define SLA thresholds that automatically mark a test as passed or failed.
+
+## Configuration
+
+```yaml
+assertions:
+  - metric: p95_latency
+    operator: less_than
+    value: 500
+    enabled: true
+```
+
+## Available metrics
+
+| Metric | Description |
+|--------|-------------|
+| `p50_latency` | 50th percentile latency (ms) |
+| `p90_latency` | 90th percentile latency (ms) |
+| `p95_latency` | 95th percentile latency (ms) |
+| `p99_latency` | 99th percentile latency (ms) |
+| `avg_latency` | Average latency (ms) |
+| `max_latency` | Maximum latency (ms) |
+| `rps` | Requests per second |
+| `error_rate` | Percentage of failed requests |
+| `success_rate` | Percentage of successful requests |
+| `total_requests` | Total requests made |
+| `total_errors` | Total failed requests |
+
+## Available operators
+
+| Operator | Meaning |
+|----------|---------|
+| `less_than` | metric < value |
+| `less_than_or_equal` | metric <= value |
+| `greater_than` | metric > value |
+| `greater_than_or_equal` | metric >= value |
+| `equal` | metric == value |
+
+## Example
+
+```yaml
+assertions:
+  - metric: p95_latency
+    operator: less_than
+    value: 500
+    enabled: true
+
+  - metric: p99_latency
+    operator: less_than
+    value: 1000
+    enabled: true
+
+  - metric: error_rate
+    operator: less_than
+    value: 1
+    enabled: true
+
+  - metric: rps
+    operator: greater_than
+    value: 100
+    enabled: true
+```
+
+Use `enabled: false` to define an assertion without enforcing it — useful for tracking a metric without failing the test.
diff --git a/docs/configuration/auth.md b/docs/configuration/auth.md
new file mode 100644
index 0000000..338837d
--- /dev/null
+++ b/docs/configuration/auth.md
@@ -0,0 +1,46 @@
+# Auth
+
+Authentication can be set per step in the scenario file.
+
+## Basic auth
+
+```yaml
+steps:
+  - method: GET
+    url: /secure
+    auth:
+      basic:
+        username: admin
+        password: secret
+```
+
+## Bearer token
+
+```yaml
+steps:
+  - method: GET
+    url: /secure
+    auth:
+      bearer: your-token-here
+```
+
+## Custom header
+
+```yaml
+steps:
+  - method: GET
+    url: /secure
+    auth:
+      header:
+        key: X-API-Key
+        value: your-api-key
+```
+
+Or use it for any arbitrary header-based auth scheme:
+
+```yaml
+auth:
+  header:
+    key: Authorization
+    value: "Token abc123"
+```
diff --git a/docs/configuration/load-profiles.md b/docs/configuration/load-profiles.md
new file mode 100644
index 0000000..b505faf
--- /dev/null
+++ b/docs/configuration/load-profiles.md
@@ -0,0 +1,77 @@
+# Load Profiles
+
+A load profile controls how worker concurrency changes over time.
+
+## constant
+
+Fixed number of workers for the full duration. Use this for steady-state benchmarks.
+
+```yaml
+load:
+  profile: constant
+  workers: 20
+  duration: 60s
+```
+
+---
+
+## ramp
+
+Linearly increases workers from `start_workers` to `end_workers` over the ramp duration. Use this to find the degradation point without shocking the system.
+
+```yaml
+load:
+  profile: ramp
+  duration: 60s
+  ramp_up:
+    start_workers: 1
+    end_workers: 50
+    duration: 30s
+```
+
+---
+
+## step
+
+Adds workers in fixed increments at regular intervals. Use this to observe how latency changes as load increases in discrete steps.
+
+```yaml
+load:
+  profile: step
+  duration: 60s
+  step:
+    start_workers: 5
+    step_size: 5
+    step_duration: 10s
+    max_workers: 50
+```
+
+---
+
+## spike
+
+Runs a base level of workers with periodic traffic bursts. Use this to test how your system recovers after a sudden surge.
+
+```yaml
+load:
+  profile: spike
+  duration: 60s
+  spike:
+    base_workers: 10
+    spike_workers: 50
+    spike_duration: 5s
+    spike_every: 15s
+```
+
+---
+
+## Stopping conditions
+
+Use `duration` to run for a fixed time, or `max_requests` to stop after a total request count:
+
+```yaml
+load:
+  profile: constant
+  workers: 10
+  max_requests: 1000
+```
diff --git a/docs/configuration/scenario.md b/docs/configuration/scenario.md
new file mode 100644
index 0000000..ba76255
--- /dev/null
+++ b/docs/configuration/scenario.md
@@ -0,0 +1,88 @@
+# Scenario File
+
+LoadForge scenarios are defined in YAML. A scenario file describes the requests to send, how to authenticate, and how much load to apply.
+
+## Full reference
+
+```yaml
+name: Test Name
+base_url: https://api.example.com
+
+scenarios:
+  - name: Scenario Name
+    weight: 100              # Relative probability when multiple scenarios are defined
+    steps:
+      - name: Step Name
+        method: GET           # GET, POST, PUT, PATCH, DELETE, HEAD
+        url: /path            # Relative (uses base_url) or absolute URL
+
+        headers:
+          X-Custom-Header: value
+
+        body:
+          raw: "raw string body"
+          json:
+            key: value
+          form:
+            field: value
+
+        auth:
+          basic:
+            username: user
+            password: pass
+          bearer: your-token
+          header:
+            key: Authorization
+            value: Bearer your-token
+
+        options:
+          timeout: 10s
+          follow_redirects: true
+          tls_skip_verify: false
+          http2: true
+
+        think: 100ms           # Pause after this step
+
+load:
+  profile: constant            # constant | ramp | step | spike
+  workers: 10
+  duration: 30s
+
+assertions:
+  - metric: p95_latency
+    operator: less_than
+    value: 500
+    enabled: true
+```
+
+## Multiple scenarios
+
+When multiple scenarios are defined, workers pick one at random weighted by the `weight` field.
+
+```yaml
+scenarios:
+  - name: Read path
+    weight: 80
+    steps:
+      - method: GET
+        url: /items
+
+  - name: Write path
+    weight: 20
+    steps:
+      - method: POST
+        url: /items
+        body:
+          json:
+            name: test
+```
+
+## Body types
+
+Only one body type is used per step. Priority: `json` → `form` → `raw`.
+
+| Type | Content-Type set automatically |
+|------|-------------------------------|
+| `json` | `application/json` |
+| `form` | `application/x-www-form-urlencoded` |
+| `raw` | none (set manually via `headers`) |
diff --git a/docs/index.md b/docs/index.md
new file mode 100644
index 0000000..0e82789
--- /dev/null
+++ b/docs/index.md
@@ -0,0 +1,49 @@
+# LoadForge
+
+> Developer-first HTTP load testing — run from the terminal or a browser-based UI.
+
+LoadForge is an open-source HTTP load testing tool built for developers and QA engineers who want to stress-test their APIs without fighting complex configuration or expensive SaaS platforms.
+
+Write a simple YAML scenario, point it at your API, and LoadForge hammers it with concurrent workers while giving you live metrics in a clean terminal UI or a browser dashboard.
+
+---
+
+## Why LoadForge?
+
+Most load testing tools are either too heavy (JMeter, Gatling) or too limited (ab, wrk). LoadForge sits in the middle:
+
+- **Zero dependencies** — a single binary, no JVM, no Node, no Docker required
+- **Two interfaces** — use the CLI for scripts and CI pipelines, or the web UI for interactive testing
+- **OpenAPI / Swagger support** — point at an API spec URL and LoadForge generates and runs the scenario for you
+- **Multiple load profiles** — constant, ramp, step, and spike traffic patterns out of the box
+- **Assertions & SLA thresholds** — define pass/fail criteria and catch regressions automatically
+- **History tracking** — every run is persisted so you can compare results over time
+
+---
+
+## Features
+
+| Feature | Description |
+|---|---|
+| **Web UI** | Browser-based dashboard to trigger, monitor, and review tests |
+| **OpenAPI / Swagger import** | Auto-generate scenarios from OpenAPI 3.x and Swagger 2.0 specs |
+| **HAR file support** | Convert or replay browser HAR recordings as load tests |
+| **Load profiles** | `constant`, `ramp`, `step`, `spike` — choose how traffic behaves |
+| **Assertions** | Define SLA thresholds (p95 latency, error rate, RPS) with pass/fail results |
+| **Auth support** | Basic auth, Bearer tokens, and custom header auth per request step |
+| **Variable injection** | Pass variables via `--var` flags or a `.env` file |
+| **History** | Persistent run history with per-request breakdown |
+| **Password protection** | Secure the web UI with bcrypt-hashed credentials |
+| **Multi-platform** | Linux, macOS, and Windows — amd64 and arm64 |
+
+---
+
+## Get started
+
+```bash
+curl -fsSL https://github.com/farhapartex/loadforge/releases/latest/download/install.sh | sudo bash
+```
+
+Then open [http://localhost:8090](http://localhost:8090) — the web UI starts automatically.
+
+See the [Installation](installation.md) guide for full details.
diff --git a/docs/installation.md b/docs/installation.md
new file mode 100644
index 0000000..a2a357f
--- /dev/null
+++ b/docs/installation.md
@@ -0,0 +1,39 @@
+# Installation
+
+## Linux / macOS (recommended)
+
+```bash
+curl -fsSL https://github.com/farhapartex/loadforge/releases/latest/download/install.sh | sudo bash
+```
+
+This single command:
+
+1. Downloads the latest `loadforge` and `loadforge-web` binaries to `/usr/local/bin/`
+2. Creates `~/.loadforge/` with a default configuration
+3. Registers and starts the web UI as a background service
+
+Open [http://localhost:8090](http://localhost:8090) immediately — no extra commands needed.
+
+Default credentials: **admin / admin**
+
+---
+
+## Manual download
+
+Pre-built binaries are available on the [releases page](https://github.com/farhapartex/loadforge/releases) for:
+
+| OS | Architecture |
+|---|---|
+| Linux | amd64, arm64 |
+| macOS | amd64, arm64 |
+| Windows | amd64 |
+
+Download the binary for your platform, make it executable, and move it to a directory on your `PATH`.
+
+---
+
+## Verify installation
+
+```bash
+loadforge version
+```
diff --git a/docs/quickstart.md b/docs/quickstart.md
new file mode 100644
index 0000000..2d6eb7c
--- /dev/null
+++ b/docs/quickstart.md
@@ -0,0 +1,50 @@
+# Quick Start
+
+## Web UI
+
+After installation the web UI starts automatically. Open [http://localhost:8090](http://localhost:8090) in your browser.
+
+Default credentials: **admin / admin**
+
+1. Paste an OpenAPI or Swagger spec URL into the input field
+2. Set your load parameters (workers, duration, profile)
+3. Click **Start Test**
+
+Live logs stream to the dashboard as the test runs. Past runs appear under **History**.
+
+---
+
+## CLI
+
+### 1. Write a scenario file
+
+```yaml title="scenario.yaml"
+name: My API Test
+base_url: https://api.example.com
+
+scenarios:
+  - name: Get users
+    steps:
+      - name: List users
+        method: GET
+        url: /users
+
+load:
+  profile: constant
+  workers: 10
+  duration: 30s
+```
+
+### 2. Run the test
+
+```bash
+loadforge run scenario.yaml
+```
+
+A real-time terminal UI shows live metrics. Press `q` to stop early.
+
+### 3. Validate without running
+
+```bash
+loadforge validate scenario.yaml
+```
diff --git a/docs/uninstallation.md b/docs/uninstallation.md
new file mode 100644
index 0000000..54fb0e8
--- /dev/null
+++ b/docs/uninstallation.md
@@ -0,0 +1,11 @@
+# Uninstallation
+
+```bash
+sudo loadforge --uninstall
+```
+
+This single command:
+
+1. Stops and removes the background service (systemd on Linux, launchd on macOS)
+2. Removes `/usr/local/bin/loadforge` and `/usr/local/bin/loadforge-web`
+3. Deletes `~/.loadforge/` including config, run history, and logs
diff --git a/docs/web-ui.md b/docs/web-ui.md
new file mode 100644
index 0000000..5d6f037
--- /dev/null
+++ b/docs/web-ui.md
@@ -0,0 +1,71 @@
+# Web UI
+
+The web UI is a browser-based dashboard for running and reviewing load tests without using the CLI.
+
+## Accessing the UI
+
+After installation the web UI starts automatically as a background service. Open [http://localhost:8090](http://localhost:8090).
+
+Default credentials: **admin / admin**
+
+Change the credentials in `~/.loadforge/web.yml`:
+
+```yaml
+username: yourname
+password: yourpassword
+```
+
+Then restart the service:
+
+=== "macOS"
+    ```bash
+    sudo launchctl unload /Library/LaunchDaemons/com.loadforge.web.plist
+    sudo launchctl load /Library/LaunchDaemons/com.loadforge.web.plist
+    ```
+
+=== "Linux"
+    ```bash
+    sudo systemctl restart loadforge-web
+    ```
+
+---
+
+## Running a test
+
+1. Paste an OpenAPI or Swagger spec URL into the input field, or select a saved scenario
+2. Set load parameters: profile, workers, duration
+3. Click **Start Test**
+
+Live logs stream to the page via Server-Sent Events as the test runs.
+
+---
+
+## History
+
+Every completed run is saved to `~/.loadforge/load_forge_history.json` (up to 100 entries). The **History** page lists all past runs with:
+
+- Latency percentiles (P50, P90, P95, P99)
+- RPS, error rate, success rate
+- Status code distribution
+- Assertion pass/fail results
+
+---
+
+## Configuration file
+
+The web server reads `~/.loadforge/web.yml`:
+
+```yaml
+addr: :8090
+username: admin
+password: admin
+session_ttl: 24h
+log_file: /home/user/.loadforge/loadforge.logs
+history_file: /home/user/.loadforge/load_forge_history.json
+
+assertions:
+  - metric: p95_latency
+    operator: less_than
+    value: 500
+    enabled: true
+```
diff --git a/mkdocs.yml b/mkdocs.yml
new file mode 100644
index 0000000..a401feb
--- /dev/null
+++ b/mkdocs.yml
@@ -0,0 +1,44 @@
+site_name: LoadForge
+site_description: Developer-first HTTP load testing — run from the terminal or a browser-based UI
+site_url: https://loadforge.readthedocs.io
+repo_url: https://github.com/farhapartex/loadforge
+repo_name: farhapartex/loadforge
+edit_uri: edit/main/docs/
+
+theme:
+  name: material
+  palette:
+    scheme: slate
+    primary: indigo
+    accent: indigo
+  features:
+    - navigation.sections
+    - navigation.top
+    - content.code.copy
+
+nav:
+  - Home: index.md
+  - Installation: installation.md
+  - Quick Start: quickstart.md
+  - Configuration:
+      - Scenario File: configuration/scenario.md
+      - Load Profiles: configuration/load-profiles.md
+      - Assertions: configuration/assertions.md
+      - Auth: configuration/auth.md
+  - Web UI: web-ui.md
+  - CLI Reference: cli.md
+  - Uninstallation: uninstallation.md
+
+markdown_extensions:
+  - admonition
+  - pymdownx.highlight:
+      anchor_linenums: true
+  - pymdownx.superfences
+  - pymdownx.tabbed:
+      alternate_style: true
+  - tables
+  - toc:
+      permalink: true
+
+plugins:
+  - search
diff --git a/scripts/install.sh b/scripts/install.sh
index 8e3301f..000f920 100755
--- a/scripts/install.sh
+++ b/scripts/install.sh
@@ -219,7 +219,7 @@ remove_service() {
 do_install() {
   require curl
 
-  local os arch version binary_name web_binary_name download_url web_download_url tmpdir
+  local os arch version binary_name web_binary_name download_url web_download_url
 
   os=$(detect_os)
   arch=$(detect_arch)