Apache-2.0 software
Free
Ongoing; optional model costs separate
- Included scope
- Local memory, MCP and dashboard
PMB
● Identity checkedLocal memory for AI coding agents · Official-source review 2026-10-09
PMB stores project facts, decisions and lessons in local SQLite/LanceDB data and exposes memory through MCP. Hooks can recall context before an agent turn and journal subsequent work.
- Category
- Local memory for AI coding agents
- Free access
- Free Apache-2.0 open-source software; optional external model and hosting costs remain separate.
- API access
- MCP stdio, CLI and local dashboard; advertised wiring for Claude Code, Cursor, Codex, Zed and other MCP clients.
Is PMB right for you?
A good fit for
Relevant to developers retaining and inspecting project memory across agents without relying on a hosted memory service.
Before you choose
Python on macOS/Linux/Windows. Recall path is described as local without LLM/network calls; optional summaries and graph extraction use the configured local or hosted model. A hosted agent may process recalled context even when the memory store is local. Follow-rate scoring and retrieval benchmarks are provider measurements, not proof of correct memory or rule enforcement. No package installed, hooks configured, memory queried or benchmark reproduced.
Workflow and handoff
Use a test workspace, inspect captured facts and hooks, and remove stale or sensitive memory before granting broader agent access.
Pricing and practical limits
Python on macOS/Linux/Windows. Recall path is described as local without LLM/network calls; optional summaries and graph extraction use the configured local or hosted model. A hosted agent may process recalled context even when the memory store is local. Follow-rate scoring and retrieval benchmarks are provider measurements, not proof of correct memory or rule enforcement. No package installed, hooks configured, memory queried or benchmark reproduced.
Review method
Official product evidence reviewed on 2026-10-09; product artwork inspected visually. No hands-on vendor application test was performed. Application performance, security claims and plan enforcement were not independently tested.
What it can do
Features & capabilities
Unknown is different from unavailable. Each fact carries its own evidence.
| Capability | Value | Evidence | Checked |
|---|---|---|---|
| Product workflow | PMB stores project facts, decisions and lessons in local SQLite/LanceDB data and exposes memory through MCP. Hooks can recall context before an agent turn and journal subsequent work. | Facts sourced | 2026-10-09 |
| Pricing and restrictions | Python on macOS/Linux/Windows. Recall path is described as local without LLM/network calls; optional summaries and graph extraction use the configured local or hosted model. A hosted agent may process recalled context even when the memory store is local. Follow-rate scoring and retrieval benchmarks are provider measurements, not proof of correct memory or rule enforcement. No package installed, hooks configured, memory queried or benchmark reproduced. | Facts sourced | 2026-10-09 |
Understand the total cost
PMB pricing & plans
Alternatives to PMB
View all ↗
chat-recallchat-recall stores coding conversations for full-text and knowledge-graph retrieval. It offers MCP access, analytics and code-related findings, with hosted and self-hosted deployment options.
Explore tool
scrittyscritty is a terminal emulator that captures attributable turns from supported coding CLIs into a searchable local corpus. Agents can query shared memory through MCP, while people can search through the terminal or CLI.
Explore tool
Aura by NaridonAura provides an open-source desktop, CLI and MCP workspace for coding agents. It advertises semantic change history, plain-language diffs, function-level rewind and intent checks, with paid synchronization and collaboration features.
Explore toolThe practical questions
Frequently asked questions
Was the application hands-on tested?
No. This listing is based on dated official-source evidence and a visual artwork check. Unconfirmed costs and restrictions are identified explicitly.
What is PMB used for?
PMB stores project facts, decisions and lessons in local SQLite/LanceDB data and exposes memory through MCP. Hooks can recall context before an agent turn and journal subsequent work.
Does PMB have a free plan?
Free Apache-2.0 open-source software; optional external model and hosting costs remain separate.. This record lists ongoing free access; check the plan limits before starting.
How much does PMB cost?
No paid monthly price is listed; this record treats the product as free to start. See the plan cards for entitlements, billing commitments and seat minimums.
Can I use PMB through an API?
MCP stdio, CLI and local dashboard; advertised wiring for Claude Code, Cursor, Codex, Zed and other MCP clients.. API access and subscription access may have different terms; consult the linked sources.
What should I check before choosing it?
Python on macOS/Linux/Windows. Recall path is described as local without LLM/network calls; optional summaries and graph extraction use the configured local or hosted model. A hosted agent may process recalled context even when the memory store is local. Follow-rate scoring and retrieval benchmarks are provider measurements, not proof of correct memory or rule enforcement. No package installed, hooks configured, memory queried or benchmark reproduced.
Price history
No retained pricing changes yet. A current price alone does not establish a historical trend.
How this profile is supported
Official product information
Current official product source read individually; product artwork inspected visually.
Read original source ↗Pricing review
Official product page reviewed for disclosed costs; unverified prices and allowances are identified in the listing.
Read original source ↗Facts apply to the named version and check date. Send a sourced correction if something changed.