Compresr is an API and SDK designed for compressing the context provided to large language models (LLMs), with the goal of reducing token usage, cost, and latency. The tool focuses on question-aware compression, meaning it retains only those parts of the input context that are relevant to a specific query, dropping extraneous tokens that do not impact the answer. This approach is positioned as an improvement over traditional methods such as trimming, truncation, or generic summarization, which can lead to loss of important information or decreased accuracy on long documents.
Users can deploy Compresr in two main ways: as a hosted SDK or as an on-premises solution. The hosted SDK is available via TypeScript and Python clients, requiring installation and an API key, and is suitable for integration into workflows where prompts or documents are compressed before being sent to an LLM. The on-premises deployment allows organizations to run Compresr within their own infrastructure, ensuring that data does not leave their network. This option is tailored for enterprise, finance, healthcare, or regulated workloads, and includes custom throughput, latency service-level agreements, and dedicated support.
Compresr offers transparent per-million-token pricing, with $10 in free credits available upon sign-up and no credit card required. The tool demonstrates significant reductions in token count and cost, with examples showing up to 90% bill reduction and compression ratios as high as 226× on long, sparse contexts. Independent benchmarks, such as FinanceBench, indicate that at light compression levels (approximately 2×), Compresr can match or exceed full-context accuracy while reducing costs by nearly half. When higher compression is prioritized, cost savings increase, though with a potential trade-off in peak accuracy.
This tool is intended for teams and organizations seeking to optimize their use of LLMs by minimizing redundant context, thereby reducing operational expenses and improving response times. It is suitable for integration with various agent harnesses and environments, including Claude Code and Cursor, and supports workflows involving long documents in formats such as PDF, Markdown, or plain text.
compresr is an AI & ML product. It focuses on reducing prompt size and optimizing token usage for large language model applications. It is built as an open-source project for LLM developers and AI engineers. compresr is open source under the Apache-2.0 license. It runs on the web, the command line, and API.
It is developed by Compresr-ai, and the product first shipped in 2026. Development happens publicly on GitHub with 8 commits in the last 90 days. Key capabilities include prompt compression, token optimization, and Python SDK. It exposes integrations via a public API.
Latest indexed changes and source events
Other apps tracked under the same category.
