Tigerless agent-memory review
Open-source, self-hosted agent memory with Markdown storage and local retrieval.
Reviewed by iTechGuides Editors · Editorial team · Updated Oct 2026
Tigerless agent-memory is an open-source runtime for giving AI agents persistent, long-term memory on a self-hosted machine. Markdown files remain the source of truth, while a rebuildable SQLite index supports ranked local retrieval. The design covers session-start injection, boundary-time writes, consolidation, and sleep-time management, with validity intervals, superseding, merging, and deletion for lifecycle control. It is aimed at solo developers and small teams that want inspectable files and local operation rather than a hosted memory service.
There is no paid tier: the project is open source with a free plan. Its retrieval stack uses BM25 through FTS5 and can add vector retrieval with reciprocal-rank fusion. Command-line and stdio MCP interfaces connect the same store to Claude Code, Codex CLI, Muse Code, Hermes, and other shell-capable agents. This shared-store approach fits workflows that move between several coding assistants while keeping memory in one local repository. No API key is required, which keeps the runtime focused on local infrastructure.
Setup is oriented toward technical users: installation is from source with Python and uv, followed by self-hosted operation. Documentation is the stated support channel, so teams looking for managed onboarding or service-level assistance may prefer a hosted alternative. The strongest case for Tigerless is a developer who wants direct control over Markdown memory, configurable retention, and explicit proposal-based management. It is less suitable for organizations that need vendor-managed deployment, centralized administration, or support beyond project documentation.
Tigerless agent-memory pros and cons
- Where it wins
- Markdown remains the canonical, portable memory store
- BM25 retrieval plus optional vector fusion runs locally
- Lifecycle hooks and proposal workflows manage memory over time
- Where it doesn't
- Self-hosted deployment requires installation from source with Python and uv
- Support is provided through documentation
- Local-only deployment may not suit teams seeking a hosted service
Tigerless agent-memory fact sheet, pricing and score →
Advertiser disclosure: iTechGuides is reader-supported. Vendors can pay for top positions in our rankings and for a place on other products' pages, and we may earn a commission when you click some links. How we rank.
Last updated · How we research and update