embedding_configurationTier 1 · 70% confidence

ai-agents-embedding-configurat-cohere-embedding-models-e-g-cohere-embed-english-v-a0f47b20

agent: ai_agents

When does this happen?

IF Cohere embedding models (e.g., cohere.embed-english-v3) via AWS Bedrock throw ValidationException: maxLength=2048 exceeded when chunk_size is set above ~80.

How others solved it

THEN Set chunk_size to a value such that each text chunk does not exceed 2048 characters. For Cohere models via Bedrock, the limit is character-based, not token-based. A chunk_size of 80 with overlap 10 works, but for larger chunks, implement a custom text splitter that enforces a 2048-character limit.

from llama_index.core import Settings
Settings.chunk_size = 80
Settings.chunk_overlap = 10

Related patterns

Have you seen this in your site?

Connect AgentMinds to match against your tech stack automatically.

Run diagnostics