Skip to content

Commit 4fdd584

Browse files
authored
Merge pull request #325 from odilitime/main
Update Quickstart Guide
2 parents d279282 + a61d63e commit 4fdd584

File tree

1 file changed

+78
-15
lines changed

1 file changed

+78
-15
lines changed

docs/docs/quickstart.md

+78-15
Original file line numberDiff line numberDiff line change
@@ -8,26 +8,28 @@ sidebar_position: 2
88

99
Before getting started with Eliza, ensure you have:
1010

11-
- [Node.js 22+](https://docs.npmjs.com/downloading-and-installing-node-js-and-npm)
11+
- [Node.js 23.1.0](https://docs.npmjs.com/downloading-and-installing-node-js-and-npm)
1212
- [pnpm](https://pnpm.io/installation)
1313
- Git for version control
14-
- A code editor (VS Code recommended)
15-
- CUDA Toolkit (optional, for GPU acceleration)
14+
- A code editor ([VS Code](https://code.visualstudio.com/) or [VSCodium](https://vscodium.com) recommended)
15+
- [CUDA Toolkit](https://developer.nvidia.com/cuda-toolkit) (optional, for GPU acceleration)
1616

1717
## Installation
1818

1919
1. **Clone and Install**
2020

21+
Please be sure to check what the [latest available stable version tag](https://github.com/ai16z/eliza/tags) is.
22+
2123
```bash
2224
# Clone the repository
2325
git clone https://github.com/ai16z/eliza.git
26+
# Enter directory
2427
cd eliza
28+
# Switch to tagged release
29+
git checkout v0.0.10
2530

2631
# Install dependencies
2732
pnpm install
28-
29-
# Install optional Sharp package if needed
30-
pnpm install --include=optional sharp
3133
```
3234

3335
2. **Configure Environment**
@@ -55,21 +57,34 @@ Eliza supports multiple AI models:
5557
- **Grok**: Set `XAI_MODEL=grok-beta`
5658
- **OpenAI**: Set `XAI_MODEL=gpt-4o-mini` or `gpt-4o`
5759

58-
For local inference:
60+
You set which model to use inside the character JSON file
61+
62+
### Local inference
63+
64+
#### For llama_local inference:
5965

6066
1. Set `XAI_MODEL` to your chosen model
6167
2. Leave `X_SERVER_URL` and `XAI_API_KEY` blank
6268
3. The system will automatically download the model from Hugging Face
69+
4. `LOCAL_LLAMA_PROVIDER` can be blank
70+
71+
Note: llama_local requires a GPU, it currently will not work with CPU inference
72+
73+
#### For Ollama inference:
74+
75+
- If `OLLAMA_SERVER_URL` is left blank, it defaults to `localhost:11434`
76+
- If `OLLAMA_EMBEDDING_MODE` is left blank, it defaults to `mxbai-embed-large`
6377

6478
## Create Your First Agent
6579

66-
1. **Edit the Character File**
67-
Check out `src/core/defaultCharacter.ts` to customize your agent's personality and behavior.
80+
1. **Create a Character File**
81+
Check out `characters/trump.character.json` or `characters/tate.character.json` as a template you can use to copy and customize your agent's personality and behavior.
82+
Additionally you can read `packages/core/src/defaultCharacter.ts`
6883

69-
You can also load custom characters:
84+
You can also load a specific characters only:
7085

7186
```bash
72-
pnpm start --characters="path/to/your/character.json"
87+
pnpm start --character="characters/trump.character.json"
7388
```
7489

7590
2. **Start the Agent**
@@ -95,7 +110,7 @@ Add to your `.env`:
95110
TWITTER_USERNAME= # Account username
96111
TWITTER_PASSWORD= # Account password
97112
TWITTER_EMAIL= # Account email
98-
TWITTER_COOKIES= # Account cookies
113+
TWITTER_COOKIES= # Account cookies (auth_token and CT0)
99114
```
100115

101116
### Telegram Bot
@@ -123,20 +138,20 @@ npx --no node-llama-cpp source download --gpu cuda
123138

124139
```bash
125140
# Start chat interface
126-
pnpm run shell
141+
pnpm start
127142
```
128143

129144
### Run Multiple Agents
130145

131146
```bash
132-
pnpm start --characters="agent1.json,agent2.json"
147+
pnpm start --characters="characters/trump.character.json,characters/tate.character.json"
133148
```
134149

135150
## Common Issues & Solutions
136151

137152
1. **Node.js Version**
138153

139-
- Ensure Node.js 22+ is installed
154+
- Ensure Node.js 23.1.0 is installed
140155
- Use `node -v` to check version
141156
- Consider using [nvm](https://github.com/nvm-sh/nvm) to manage Node versions
142157

@@ -152,6 +167,54 @@ pnpm start --characters="agent1.json,agent2.json"
152167
- Check GPU compatibility
153168
- Ensure proper environment variables are set
154169

170+
4. **Exit Status 1**
171+
If you see
172+
```
173+
ERR_PNPM_RECURSIVE_RUN_FIRST_FAIL @ai16z/agent@0.0.1 start: node --loader ts-node/esm src/index.ts "--isRoot"
174+
Exit status 1
175+
ELIFECYCLE Command failed with exit code 1.
176+
```
177+
178+
You can try these steps, which aim to add `@types/node` to various parts of the project
179+
180+
```
181+
# Add dependencies to workspace root
182+
pnpm add -w -D ts-node typescript @types/node
183+
184+
# Add dependencies to the agent package specifically
185+
pnpm add -D ts-node typescript @types/node --filter "@ai16z/agent"
186+
187+
# Also add to the core package since it's needed there too
188+
pnpm add -D ts-node typescript @types/node --filter "@ai16z/eliza"
189+
190+
# First clean everything
191+
pnpm clean
192+
193+
# Install all dependencies recursively
194+
pnpm install -r
195+
196+
# Build the project
197+
pnpm build
198+
199+
# Then try to start
200+
pnpm start
201+
```
202+
5. **Better sqlite3 was compiled against a different Node.js version**
203+
If you see
204+
205+
```
206+
Error starting agents: Error: The module '.../eliza-agents/dv/eliza/node_modules/better-sqlite3/build/Release/better_sqlite3.node'
207+
was compiled against a different Node.js version using
208+
NODE_MODULE_VERSION 131. This version of Node.js requires
209+
NODE_MODULE_VERSION 127. Please try re-compiling or re-installing
210+
```
211+
212+
You can try this, which will attempt to rebuild better-sqlite3.
213+
214+
```bash
215+
pnpm rebuild better-sqlite3
216+
```
217+
155218
## Next Steps
156219

157220
Once you have your agent running, explore:

0 commit comments

Comments
 (0)