Introducing Chroma Context-1, a 20B parameter search agent.
> pushes the pareto frontier of agentic search
> order of magnitude faster
> order of magnitude cheaper
> Apache 2.0, open-source
Introducing our latest technical report: Context Rot - How Increasing Input Tokens Impacts LLM Performance
Our results reveal that models do not use their context uniformly.
full report in replies
We introduce representative generative benchmarkingβcustom eval sets built from your own data that reflect real user queries.
thank you @wandb for collaborating!
link to report in replies
Introducing Chroma Package Search
Enable your AI agents to search the source code of your package dependencies
Includes 3 MCP tools:
package_search_hybrid()
package_search_grep()
package_search_read_file()
Add it to Cursor, Claude Code, VS Code, and Codex in 5 seconds
A deep dive on wal3 - Chroma's open-source write-ahead log built on object storage.
Featuring:
- a 30-year-old lock-free algorithm
- Amazon S3's newest conditional writes feature
- a novel checksumming technique called setsum
We all have access to the same models, what coding agent you use matters.
For example, @FactoryAI with Claude Sonnet 4 can out perform Claude Code with Claude Opus 4.
Learn about how Chroma Cloud plays a role in Factory's context engineering strategy: trychroma.com/company/factorβ¦
Prompting isnβt engineering, argues @isaacbmiller1. βContext engineering is how we actually build reliable AI systems.β Learn how @DSPyOSS turns prompts into programs.