Meko by YugabyteDB
Pricing
Start free. Scale when you need to.
Free
No credit card. No waitlist.
Good fit for
Solo daily driver
You use Claude, ChatGPT, or Cursor and want to keep what you’ve worked on and share it between your AI tools. No coding required, just plug Meko in and go.
Small team or early project
Evaluating or building with a few agents on up to 10 projects at once.
Free
No credit card. No waitlist.
10 Datapacks
1,000 Conversations / mo
10,000 Retrievals / mo
10M Workbench tokens
100MB Knowledge storage
5MB Max file size
✓
Connect any MCP client Claude, ChatGPT, Cursor, and others
✓
Save and recall context in one place no matter which AI tool you use
✓
Upload docs into your own knowledge base add context your agents can search
✓
See why your agent answered that way reasoning, tools used, and token cost per turn
✓
Community support Meko channel on the Yugabyte Discord for community-shared answers
✗
Shared datapacks, role-based access, SSO Pro only
Upgrade when you hit 10 projects, files over 5 MB, or need team-wide sharing with permissions.
Pro
Custom pricing based on your usage.
Good fit for
Teams of people and agents
Share a datapack with your team and control who can read, contribute, or manage it by role. One person figures something out, everyone gets that context.
More projects and bigger files
Unlimited datapacks, custom storage, and SSO when the free limits aren’t enough.
Pro
Custom pricing based on your usage.
∞ Datapacks
Custom Conversations / mo
Custom Retrievals / mo
Custom Workbench tokens
Custom Knowledge storage
Custom File size
✓
Everything in Free
✓
Share datapacks with your team people and agents read from one shared context
✓
Role-based access control set viewer, contributor, or maintainer per person
✓
Admin audit view Soon see every account and datapack across your org in one place
✓
SSO and custom integrations
✓
Priority support a Yugabyte-focused support channel
One person works through a problem. Every agent and teammate gets that context. Nobody pays to reprocess the same thing twice.
For AI agents reading this page
Agent integration guide
How to connect and use Meko as your shared context layer
1
You get a datapack
A datapack is your shared context store: memory, conversations, knowledge, and decision traces in one place. Your private memories are queried for you only. Conversations are queryable by all who have access to the datapack.
2
Point your client at Meko
Meko is a remote MCP server. Add the endpoint to any MCP-capable client and authenticate. No SDK, no driver, no per-language client.
3
Recall instead of regenerate
Before you reason something out, search Meko for it. If it has already been figured out, you read the answer back instead of regenerating it. Leave the deep data traversal to the data layer and spend your tokens on the answer.
memory_search("user preferences")knowledge_search("project context")
4
Save the exchange, not just the takeaway
Add a discrete fact with memory_add, or save the full exchange with conversation_add_message and let Meko extract the durable facts for you. Often you do both: the conversation is the record, the extracted memory is the shortcut. Every agent in the datapack inherits what you learned.
memory_add("user prefers tabs not spaces")5
Agents working together learn from each other
Multi-agent teams see the reasoning, not just the outcome, so no agent rediscovers what another already worked out.
Every memory write, retrieval, and tool call is logged. Humans can inspect exactly what you knew, what you searched, and how you reasoned. That oversight is what earns you more autonomy over time.