nl2uri

AI Cost Tracking

PyPI Version Python License AI Cost Human Time Model

Generated on 2026-06-15 using openrouter/qwen/qwen3-coder-next


TellMesh URI package extracted from tellmesh/tellmesh.

Role: natural language → URI plans (single, list, tree, compact flow, task graph, workflow graph).

prompt -> nl2uri plan/graph/flow/task -> uri3 validate/run -> uri2ops import-graph (operator slice)

Install

pip install -e .
# LLM planners need OpenRouter key:
set -a && source .env && set +a   # OPENROUTER_API_KEY

CLI quickstart

nl2uri plan -p "open browser and check health" --validate
nl2uri flow -p "weather agent health check" --expand --validate
nl2uri graph -p "browser health check with dom assert" --llm --validate --dry-run
nl2uri task -p "android screenshot" --llm --validate
nl2uri tree -p "weather map agent" --out output/weather.uri.tree.yaml
nl2uri generate -p "invoices agent with health" --out output/invoices.uri.tree.yaml

# Operator bridge (graph YAML -> uri2ops task):
nl2uri graph -p "..." --llm -o /tmp/graph.yaml
uri2ops import-graph /tmp/graph.yaml --validate --out /tmp/task.yaml
uri2ops run /tmp/task.yaml --adapter mock --approve

Examples

Example Path
NL → graph (LLM) tellmesh/examples/16_llm_graph_planner
NL → flow tellmesh/examples/18_llm_flow_planner
Multi-URI graph tellmesh/examples/13_nl2uri_multi_uri_graph
Operator bridge uri2ops/examples/16_nl2uri_operator_bridge

License

Licensed under Apache-2.0.