Problemsby GlyphCrewAI users keep filing bugs
9 posts from 9 people8d activeSteady, +1 this week
Posts per day
The posts behind it
[FEATURE] Add Asyntai website knowledge tools
### Feature Area Integration with external tools ### Is your feature request related to a an existing bug? Please link it here. NA ### Describe the solution you'd like Add tools for [Asyntai](https://asyntai.com) to `crewai-tools`. Asyntai is an AI chatbot platform for websites....
[BUG] ScrapflyScrapeWebsiteTool mandates api_key, breaking env var usage
### Description In `crewai-tools`, `ScrapflyScrapeWebsiteTool` defines `api_key: str` as a required positional argument in `__init__` without a default value (`= None`). However, the tool declares `SCRAPFLY_API_KEY` in `env_vars`: and internally resolves the key via: Because...
[BUG] docs: Incorrect CrewBase import in hook documentation examples
### Description ### Description The examples in the following documentation pages import `CrewBase` directly from `crewai`: - `docs/v1.15.30/en/learn/before-and-after-kickoff-hooks.mdx` - `docs/v1.15.30/en/learn/execution-hooks.mdx` However, attempting this import with CrewAI...
docs(pt-BR): broken links to pt-BR quickstart in galileo observability guide (missing leading slash)
In the pt-BR translation of the Galileo observability guide, the two CrewAI Quickstart links on lines 25 and 32 are written as relative paths without a leading slash: - docs/edge/pt-BR/observability/galileo.mdx:25 - [CrewAI Quickstart](pt-BR/quickstart) -...
[BUG] Completed A2A artifact text is omitted from the next LLM prompt
### Description When a remote A2A task completes, CrewAI extracts the reply from the task artifact into `a2a_result["result"]`, then builds the next local LLM prompt from `task.history` only. With the default `trust_remote_completion_status=False`, that `result` is not returned...
[BUG] Converter._coerce_response_to_pydantic drops agent reference on fallback retries
### Description ### Summary When model validation fails inside `Converter._coerce_response_to_pydantic()`, line 67 calls `handle_partial_json` with a hardcoded `agent=None`: If `handle_partial_json` falls through to `convert_with_instructions`, it raises: This aborts the...
[FEATURE] Add a TraceRoot observability guide
### Feature Area Documentation ### Is your feature request related to a an existing bug? Please link it here. N/A ### Describe the solution you'd like Hi, I’m Leilei from TraceRoot. We have a [CrewAI integration](https://traceroot.ai/docs/integrations/crewai) and would like to...
[Feature Request]: Add zero-auth OpenMeteoWeatherTool to crewai-tools
### Feature Area Core functionality ### Is your feature request related to a an existing bug? Please link it here. No, this is a standalone feature request to add a zero-authentication weather tool to crewai-tools. ### Describe the solution you'd like Add OpenMeteoWeatherTool to...
[BUG] Agent(max_iter=0) / negative max_iter accepted, then kickoff raises nonsensical RecursionError
### Description `Agent(max_iter=0)` and negative `max_iter` are accepted at construction. On `Crew.kickoff()`, CrewAI raises a `RecursionError` from the Flow loop detector with nonsense counts (`called 0 times` / `called -10 times`), instead of validating the bound or finishing...
Companies and products named
About this problem
Evidence
9 posts from 9 people in 1 place, about 1 a week over 43 days. Mostly on crewAIInc/crewAI.
Frustration Frustration 0 of 3· Seen on GitHub issues
How it was grouped
Posts about CrewAI that state a pain but match no known issue yet. First post Sep 25, 2026, latest Nov 6, 2026. Corroborated: 3 or more posts from 2 or more people or places. Method
Other CrewAI problems
CrewAI upgrades break working setups
3 posts from 2 people on GitHub issues.
History