AI brand PR assistant
Media coverage monitoring and PR analysis reports
PRODUCT TOUR
From topic research to publishing

01 / Topic research
AI-guided topic research
Talk to Heimdall AI to work through the context of a topic. It asks questions to sharpen your angle and cites live news material as evidence.
CORE CAPABILITIES
What Heimdall does
Multi-source integration
PR sources are structured automatically, then deduplicated and clustered by topic, so coverage across the web sits in one view. News sites, social media, forums and blogs are all supported.
Sentiment intelligence
LLM comprehension delivers precise sentiment classification and keyword-cloud visualisation. Watch the trend in brand coverage as it moves and catch crisis signals early.
AI PR drafting
Press releases, Q&As and social copy are drafted in your brand's voice. AI speeds up the whole path from monitoring to response, freeing your PR team for the creative work.
WHAT CHANGES
What changes for your team
AI sorts the topics
High-volume events are classified automatically, so nobody collates them by hand
Crisis response starts sooner
Monitoring and finished copy happen in the same platform
Focus on the real signals
Noise is filtered out, and the topics worth acting on come first
USE CASES
PR and media scenarios
- Brand crisis alerts and live monitoring
- Competitor share of voice and market intelligence
- Automated post-campaign PR impact reports
Positive sentiment 78%
Positive sentiment 65%
Positive sentiment 52%
HEIMDALL MCP SERVER
Heimdall data in Claude
The official MCP package from Asgard AI
26 MCP tools
Workspaces, articles, Avatars, missions and topics are all covered, so every core resource in Heimdall can be read. The tool count is from the README of asgard-ai-platform/mcp-heimdall on GitHub.
Read-only by design
Every tool is read-only. No write API is exposed: the assistant can read your data but never change it.
One command to integrate
Works with Claude Desktop and Claude Code. Once it is set up, you can work with Heimdall data directly inside an AI conversation.
Install the MCP package with pip, or run it directly with uvx:
pip install mcp-heimdalluvx --from mcp-heimdall mcp-heimdallexport HEIMDALL_API_TOKEN=your_bearer_token_hereasgard-ai-platform / mcp-heimdall
MIT License · Open Source
PRICING
