AsterMind Virtual Assistant
Answers from your documents, with the sources to prove it.
The AsterMind Virtual Assistant answers questions from your organization's documents and cites its sources, using AsterMind's own search and ranking together with a language model you choose, either as a hosted service or installed on your own servers.
Or try it now: the assistant in the corner of this page is the AsterMind Virtual Assistant.
How it works
- 1
Upload your documents
PDF, TXT, Markdown, CSV or JSON.
- 2
AsterMind finds the passages that answer the question
With its own search and ranking.
- 3
The language model you choose writes the answer
And the assistant lists the sources it used.
Choose your language model
OpenAI, Anthropic, AWS Bedrock, or an open model you host yourself, such as Gemma, through Ollama or vLLM.
Run it where you need it
Hosted by AsterMind
On AWS in the United States, for one organization or many.
On your own servers, including fully air-gapped
Linux, with an open model you host yourself, such as Gemma, through Ollama or vLLM. No GPU required; a GPU makes a local model faster. Docker is supported.
On your website
An embeddable chat widget. If our server can't be reached, the widget can still search content cached in the browser.
Built for teams
Roles
System administrators, tenant administrators and tenant users.
Your data, scoped to you
Every stored record is tied to its tenant.
Encrypted
The hosted service encrypts data at rest and in transit, and stored model credentials are encrypted with AES-256-GCM.
When the language model is down
A self-hosted install still returns short answers or the most relevant passages from your documents.
Pricing
Monthly plans for the AsterMind Virtual Assistant. Start small and scale to enterprise volumes.
Starter
$350 / month
- 25,000 messages / month
- ~7,000 sessions
- 1 tenant, 1 user
- Best for: Pilots and small teams
Professional
$1,499 / month
- ~175,000 messages / month
- 50,000 sessions
- 1 tenant, unlimited users
- Best for: Growing organizations
Business
$3,499 / month
- ~525,000 messages / month
- 150,000 sessions
- Up to 3 tenants, unlimited users
- Best for: High-volume teams
Enterprise
Contact us
- More than 1M messages / month
- 250,000 sessions / month
- Unlimited tenants and users
- Best for: Enterprise scale
Every plan includes
- Unlimited documents
- The embeddable chat widget
- Encryption at rest and in transit
- Automatic updates
- Bring your own model: OpenAI, Anthropic, AWS Bedrock or Ollama
- Email support
A session is one continuous conversation; a new one starts after 30 minutes idle or when the user starts over. A message is one user turn or one assistant reply. Allowances reset monthly. Overage: $0.03 per session month-to-month, $0.02 on annual prepay.