pumaDB vs Meridian: Detailed Comparison

Overview

pumaDB and Meridian are two distinct tools that address different aspects of AI and work management. pumaDB is a hosted memory service designed to give AI agents a persistent, shared place to store and retrieve JSON data, such as preferences, task state, and research notes. It integrates via MCP (Model Context Protocol) or REST, making it easy to add durable memory to agents like ChatGPT, Claude, or Codex without setting up a database.

Meridian, on the other hand, is an open-source, on-device AI work journal. It automatically captures your work activity—like coding, debugging, and meetings—and generates plain-English summaries and status updates for project management tools like Jira. It runs entirely locally, ensuring your data never leaves your device unless you explicitly approve a worklog to post.

Feature Comparison

FeaturepumaDBMeridian
Core PurposeMemory layer for AI agentsAutomatic work journal and status updater
DeploymentCloud-hosted (SaaS)On-device, open-source
Data StorageJSON rows in tables (20 tables, 1,000 rows each, 25 MB)Local journal with per-day entries
IntegrationMCP (ChatGPT, Claude, Codex) and RESTJira, Linear, GitHub Projects, Azure DevOps, Trello
Memory ManagementTools like 'remember' for typed memoryAutomatic capture and summary generation
Safety & PrivacyVersion history, rate limits, reviewable memory100% local, open-source, no data leaves device
PricingNot specified (likely paid with free tier)Free (MIT license)

Pricing

pumaDB's pricing is not explicitly listed on its website, but given its hosted nature, it likely follows a SaaS model with a free tier (limited to 20 tables, 1,000 rows per table, and 25 MB storage) and paid plans for higher limits. Users should check the official site for current pricing.

Meridian is completely free and open-source under the MIT license. There are no paid tiers, no trials, and no credit card required. You can download and use it indefinitely at no cost.

Pros and Cons

pumaDB

Pros:

  • Simple setup: no database or infrastructure management.
  • Flexible: works with any MCP-compatible agent or REST client.
  • Built-in safety: version history, rate limits, and reviewable memory.

Cons:

  • Cloud-based: data is stored on pumaDB servers, which may raise privacy concerns.
  • Limited storage: 25 MB total per account may be restrictive for heavy use.
  • Pricing unclear: no transparent pricing page.

Meridian

Pros:

  • Privacy-first: 100% local, no cloud, no account.
  • Open source: fully auditable code.
  • Automatic: captures work without manual input, drafts updates for PM tools.

Cons:

  • Limited to work journaling: not a general-purpose memory store.
  • Requires desktop app installation and screen monitoring.
  • PM tool integrations are read-only until approval, which may be limiting for some workflows.

Verdict

Choose pumaDB if you need a flexible, cloud-based memory layer for AI agents across various tools, and you're comfortable with a hosted service. Choose Meridian if you prioritize privacy and want an automatic, local work journal that integrates with PM tools, and you prefer open-source software. Both tools serve different needs, so the best choice depends on your specific workflow and privacy requirements.