llm-browser-testkit 0.2.5

LLM-driven browser test framework — define browser test scenarios in TOML with natural language steps and LLM-powered assertions
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36

# Anthropic / custom provider example — thinking disabled, model_params for provider-specific options
# Run: llm-browser-testkit run examples/custom-provider.toml \
#        --llm-url https://api.anthropic.com \
#        --llm-model claude-sonnet-4-20250514 \
#        --llm-api-key $ANTHROPIC_API_KEY

[config]
base_url = "https://example.com"
timeout_secs = 30
start_url = "/"

# For Anthropic: don't send `thinking` at all (leave unset / set to null in TOML).
# Use `model_params` to pass provider-specific fields like `effort`.
# For OpenAI with thinking: set `thinking = true` or `thinking = false`.
[config.model_params]
effort = "high"

[[definitions]]
name = "no_errors"
preset = "no_error_on_page"

[[test]]
name = "Smoke test"

[[test.steps]]
kind = "navigate"
url = "/"
wait_after_ms = 2000

[[test.steps]]
kind = "assert"
definition = "no_errors"

[[test.steps]]
kind = "screenshot"
path = "anthropic-smoke.png"