Helix AI Labs · Valence

Practical setup guides for Valence

Use these walkthroughs to choose a provider, add the right key, turn on local memory, and move past conversations into Valence without guesswork.

Let AI Walk You Through It

Copy this prompt and paste it into any AI chat (ChatGPT, Claude, Gemini) for personalized, step-by-step setup help. For complete feature documentation, share this URL with your AI: helixailabs.com/valence_docs.html

I just installed Valence, a cross-model AI chat interface for Windows. I need help getting an API key for [PROVIDER NAME]. Read the full user guide at https://helixailabs.com/valence_docs.html for context. Walk me step by step through: 1) Creating an account, 2) Finding the API key page, 3) Understanding free tier limits and costs. Keep it simple and practical.

Provider Setup Guides

OpenAI
$5 starter credit

GPT-4o, GPT-4 Turbo, and o1 reasoning models. The most widely-used AI platform.

Key format: sk-...
Free tier: $5 credit for new accounts
Get API Key ↗
Anthropic
$5 starter credit

Claude 4, Claude 3.5 Sonnet, and Claude 3 Haiku. Strong at coding, analysis, and long-context tasks.

Key format: sk-ant-...
Free tier: $5 credit for new accounts
Get API Key ↗
Google Gemini
Generous free tier

Gemini 2.0 Flash and Gemini Pro. Fast, multimodal, and has the most generous free tier of any major provider.

Key format: AI...
Free tier: Gemini Flash free with rate limits
Get API Key ↗
Azure OpenAI
$200 Azure credit

Enterprise-grade OpenAI models hosted on Azure. Requires an Azure subscription and resource deployment.

Key format: Varies by resource
Free tier: $200 credit for new Azure accounts
Open Azure Portal ↗
xAI (Grok)
$25/mo free credit

Grok-2 and Grok-3 models from xAI. Real-time knowledge and competitive reasoning performance.

Key format: xai-...
Free tier: $25/month free API credit
Get API Key ↗
OpenRouter
Some free models

Unified gateway to 100+ models from many providers. One API key, many models. Great for experimentation.

Key format: sk-or-...
Free tier: Select free models available
Get API Key ↗
💻 Ollama (Local)
100% Free • Local

Run open-source models on your own hardware. No API key needed, no usage costs, complete privacy. Requires a decent GPU for best results.

Key format: No key needed
Cost: Completely free, runs locally
  1. Download and install from ollama.com/download
  2. Open a terminal and run ollama pull llama3
  3. In Valence, select Ollama as your provider — it connects automatically

Local Memory

🧠 What Is Local Memory?
v1.1+

Local Memory gives Valence persistent memory across conversations. When enabled, the AI can remember things you've discussed and recall them later — all stored locally on your machine, never sent to Helix servers.

  1. Open Settings → Memory
  2. Toggle Enable Local Memory on
  3. Choose a Save Mode:
    Summary — AI summarizes conversations into compact memories (recommended)
    Full — saves complete message history (uses more disk space)
  4. Set Auto-Save Interval — how many AI responses between automatic saves (0 = manual only)
  5. Enable Recall Triggers to let the AI automatically search memory when you ask about past conversations
🔍 Memory Browser

Browse, search, and manage all your saved memories from Settings → Memory.

  1. Browse — memories are listed with title, date, summary, and topic tags. Only 20 load at a time; click Show More to load more.
  2. Filter — type in the search box to instantly filter memories by title, summary, or topics.
  3. Deep Search — press Enter or click Search to run an AI-powered deep search across all memory content.
  4. Delete — click the trash icon on any memory to remove it permanently.
  5. Manual Save — click Save Current Chat to Memory to save the active conversation at any time.

Importing Conversations

Import from ChatGPT
Two options

Import your entire ChatGPT history into Valence's local memory. Choose between a browser login (fastest) or a ZIP export.

  1. Open Settings → Memory → Import Conversations
  2. Option A — Browser Login (recommended): Click Import from ChatGPT. A browser window opens. Log in to your OpenAI account, then confirm the import. Valence fetches your conversations automatically via the API.
  3. Option B — ZIP Export: Click Import from ZIP file. First, export your data from ChatGPT Settings → Data Controls → Export Data (takes a few minutes to receive the email). Then select the downloaded ZIP file.
  4. Conversations are converted to local memories with titles, summaries, and topic tags.
Import from Claude

Import your Claude conversation history into Valence's local memory via browser login.

  1. Open Settings → Memory → Import Conversations
  2. Click Import from Claude
  3. A browser window opens. Log in to your Anthropic account, then confirm the import.
  4. Valence fetches your conversations and indexes them as local memories.
Import from Gemini
Slower — page by page

Import your Google Gemini conversations. Gemini doesn't offer a bulk API, so Valence navigates through your conversations one at a time.

  1. Open Settings → Memory → Import Conversations
  2. Click Import from Gemini
  3. A browser window opens. Log in to your Google account, then confirm the import.
  4. Valence reads conversations one at a time — this may take a few minutes depending on how many you have.
  5. Progress updates appear in real-time as each conversation is imported.

Ready to Chat?

Once you have an API key, open Valence → Settings → API Keys and paste it in. You'll be chatting in seconds.

Explore Valence Features