The Intelligence
Layer for your Vault.
AstraClaw is KeepCloud's AI-powered file intelligence engine. Search, summarize, and reason across your entire unified vault using natural language — powered by the AI model of your choice.
Getting Started
AstraClaw is built directly into your KeepCloud dashboard. No separate installation or extensions required. Follow these steps to get up and running.
Log in to KeepCloud
Go to dashboard.keepcloud.in and sign in to your unified workspace.
Connect a cloud account
Navigate to Integrations and securely connect Google Drive, OneDrive, or Dropbox.
Open AstraClaw AI
Click the glowing 'AstraClaw AI' button in the left sidebar.
Grant indexing consent
On first launch, grant AstraClaw permission to securely index your files for semantic search.
Wait for initial sync
AstraClaw begins scanning. You can start chatting immediately with what has been indexed.
Pro Tip: Query while indexing
You don't have to wait for 100% completion. You can start asking questions while indexing is in progress in the background. AstraClaw searches what is indexed so far and dynamically finds more as the sync completes.
Vault & Sources
AstraClaw acts as an omnipresent intelligence layer across all your connected cloud accounts by default. You also have fine-grained control over which files are used as context.
Adding a cloud account
Navigate to dashboard.keepcloud.in/integrations. Choose Google Drive, OneDrive, or Dropbox and follow the standard OAuth flow.
Multiple accounts are fully supported. You can connect multiple accounts from the same provider (e.g., your personal and work Google accounts), and AstraClaw will seamlessly index and search across all of them simultaneously.
Pinning specific sources
Need hyper-focused answers? Inside the AstraClaw chat panel, click the Your sources tab in the left sidebar, then click Add to search and attach specific files or entire folders. These pinned items are given strict priority in the AI's reasoning pipeline.
Enabling / disabling a source
Go to Intelligence Hub → Sources and toggle any connected account on or off. When an account is toggled off, its files are immediately excluded from all AI searches and vector retrievals.
Chat & Search
The AstraClaw chat panel works like a conversational assistant that has memorized every document in your vault. Ask it anything — it finds the right files, extracts relevant context, and synthesizes a direct answer.
The art of prompting
The AI will struggle because it doesn't know which of your 50 reports you mean.
The AI knows exactly what semantic vectors to retrieve and what to extract.
Attaching files directly in chat
Click the Paperclip icon directly inside the chat input. This allows you to search your vault and attach specific files to your prompt. AstraClaw will then bypass the global vault search and only read the attached files for that specific question, giving you perfectly isolated answers.
Understanding source cards
AstraClaw always cites its sources. Below each AI response, you will see interactive Source Cards representing the exact chunks of files the AI used to synthesize its answer. Click any card to instantly preview the original file.
Chat History
AstraClaw treats conversations as persistent sessions, meaning you can close the app, come back a week later, and pick up a complex analysis exactly where you left off.
How sessions work
A session represents one continuous thread of thought. When you click the Pencil icon in the top right header of the panel, a fresh session begins. All messages sent and received within that thread are securely stored together.
Viewing your history
Click the History tab in the left sidebar of the AstraClaw panel. Your past sessions are intelligently grouped by recency (Today, Yesterday, Earlier Chats). Simply click any session card to reload the entire context and resume chatting.
Performance & Cache Expiry
To ensure lightning-fast load times, AI-generated answers are temporarily cached. If you open a very old session, the original answer text may have expired from the cache layer and will display as [Answer expired]. However, your original question text and the context of the session are permanently preserved.
AI Models
AstraClaw is model-agnostic. While it comes batteries-included with KeepCloud's shared model pool, power users can configure their own API keys for higher rate limits, stricter privacy, or access to state-of-the-art reasoning models.
Adding your own provider
1. Open the Intelligence Hub (click the Settings ⚙️ icon in the AstraClaw panel header).
2. Navigate to the AI Models tab.
3. Select the capability category (Intelligence for text, Embeddings for vectors, Vision for images).
4. Click + Add Key, select your provider, paste your API key, and hit Save.
Supported providers
Cost-free AI Inference
Providers like Groq, NVIDIA NIM, Google Gemini, and DeepSeek currently offer generous free API tiers. By configuring these in AstraClaw, you can utilize enterprise-grade AI search across your vault at absolute zero cost.
Vault Sync
Before AstraClaw can magically answer questions about your files, it must process them through our secure indexing pipeline. This converts human-readable text into mathematical representations (vectors) that AI models understand.
Automatic vs. Manual Sync
KeepCloud actively monitors your connected cloud accounts. When you add or modify files in Google Drive or OneDrive, AstraClaw automatically discovers and queues them for background indexing.
If you ever feel a newly added file is missing from your search results, you can force an immediate sweep by going to Intelligence Hub → Vault Sync and clicking Re-sync.
Intelligence Hub
The Intelligence Hub is the mission control centre for your AI infrastructure. Access it via the ⚙️ icon in the AstraClaw header.
Vault Sync
Monitor real-time indexing progress, view your total indexed node count, and manually trigger full re-syncs.
AI Models
Configure custom API keys for Intelligence (LLMs), Embeddings, and Vision models to take full control of your stack.
Sources
Micro-manage your context window by toggling specific cloud accounts on or off for semantic indexing.
System Status
Live telemetry dashboard showing indexed file counts, total vector embeddings, and real-time retrieval latency.
Privacy & Security
Your data privacy is the unshakeable foundation of KeepCloud. We designed AstraClaw so you don't have to compromise security for intelligence.
Your file content is stored in your own cryptographically isolated vector space. It is strictly mathematically separated from all other users.
We strictly prohibit the use of your private files, prompts, or chat history to train global foundational AI models.
Any custom API keys you provide are encrypted at rest using AES-256 before being stored. They are never logged or exposed in plain text.
AstraClaw will never index a single byte of your files without your explicit, revocable consent. You are in total control.
Data Deletion
We respect your right to be forgotten. To permanently delete all indexed data, email support@keepcloud.in. Our automated systems will cryptographically shred all vectors, embeddings, and search history associated with your account within 48 hours.
For the complete legal framework, please visit our global Privacy Policy.