Fetch.ai
ASI1 Integration
Last updated
ASI1 Integration
Last updated
Status: Planned
Type: LLM Provider
Source:
Docs:
ASI1 is the first Web3-native large-language model built for agentic AI. It combines multi-step reasoning with native $FET wallet support, enabling secure, context-aware conversations for autonomous agents operating in decentralised environments.
Current model: asi1-mini (REST + streaming APIs)
Dynamic reasoning modes: Multi-Step, Complete, Optimised, and Short, automatically selected per request.
Agent connectivity: ASI1 can call, or be extended by, external uAgents that act as domain experts and can be monetised through Agentverse.
LLM plug-in: Drop-in ASI1 client with API-key auth
Agent-level routing: Select ASI1, OpenAI, or local models per call
Adaptive reasoning: Surface ASI1’s mode selector to let swarms trade depth vs. speed
uAgent bridge: Expose JuliaOS agents in Agentverse so ASI1 can delegate tasks back to the swarm
Wallet metering: Optional FET-wallet passthrough once live billing is enabled
Cross-chain trade optimiser – ASI1 reasons; JuliaOS executes swaps
DeFi compliance copilot – ASI1 consults legal uAgents; JuliaOS enforces rules
Research summariser – Feed lengthy PDFs to JuliaOS; ASI1 summarises and links into its knowledge graph
This integration is not yet live.
Coming deliverables (planned)
asi1 client module
Agent-level llm_provider = "asi1" configuration
Test harness & sample notebooks
Step-by-step tutorial in GitBook
Stay tuned for updates as we finalise specs and begin internal testing.