Checking an import
The platform lead asks whether last night's import into the docs collection finished. Memtro counts 48,200 points, 1,900 of them with source set to the new wiki, and shows the indexed payload keys.
Ask how many points a Qdrant collection holds, and what their payloads say.
Connect Qdrant with your cluster's host and an API key and Memtro reads collections with their size, vector settings and payload schema, and points' payloads by id, scrolled or counted under a filter. It works with Qdrant Cloud or your own server. Vectors are never fetched. When you ask, and once an administrator has granted them by name, it can upsert points, change or clear payloads, delete points, create collections, take snapshots and index payload fields. Deleting points, a collection or a snapshot is permanent, and overwrites cannot be undone from Memtro. A read-only key keeps these out of reach.
Ask in plain English from chat, your editor, a scheduled job or your own agent. The model picks the right Qdrant tool, reads what it needs, and answers with the source cited.
These tools change something in Qdrant. The model is told to use them only when you ask for exactly that, and an administrator grants each one by name to the teams that may use it; a personal connection acts only after its owner allows it.
Once Qdrant is connected, these are the kinds of things people ask. Plain questions, no syntax. If the answer lives partly in another system you have connected, Memtro looks there too.
How many points are in the support-articles collection?
Which payload keys are indexed in the products collection?
How many points have source set to confluence?
What is the payload of point 4812?
Take a snapshot of the docs collection, then delete points 101 and 102
The platform lead asks whether last night's import into the docs collection finished. Memtro counts 48,200 points, 1,900 of them with source set to the new wiki, and shows the indexed payload keys.
Dan Fletcher asks Memtro to delete two points that still describe a retired Northwind product. Memtro reads both payloads first, takes a snapshot of the collection as he asked, then deletes the two points by id, saying only that snapshot can bring them back.
The teams that lean on Qdrant most, a question each of them asks, and what they connect beside it.
A record in Qdrant is rarely the whole story. Connect the systems that hold the rest and Memtro checks them together, then remembers what it learned so the next question starts further along.
Using something else for the same job, or both? Each connector has its own tools; connect as many as you like.
On Qdrant Cloud copy a key from the cluster's API Keys tab. Self-hosted, use the read_only_api_key or the api_key from the service config. A read-only key is safer and enough for everything but the writes
Type the cluster's name in Host only, with no https://, path or port: a public DNS name serving https with a publicly trusted certificate. IP addresses, private names and plain http are refused
Paste the API key into Connections → Qdrant in Memtro with the Host, setting the port if it is not 6333
Ask how many points each collection holds
Chat in the dashboard, the OpenAI-compatible API, scheduled jobs, and any MCP client: Claude Code, OpenCode, Cursor, Claude Desktop. The Qdrant tools are the same everywhere.
Only when you ask, and only what it has been granted: upsert points by id (replaces any point that already has the id); set payload keys on points by id (overwrites those keys); delete points by id; create a collection; delete a collection and all its points (permanent); remove payload keys from points by id; replace the whole payload of points by id; clear the payload of points by id; take a snapshot of a collection; delete a collection snapshot (permanent); index a payload field; remove a payload field index. Every other tool reads, searches or lists. The Connections page labels the tools that act, an administrator grants each one by name to the teams that may use it, and a personal connection acts only after its owner allows it.
A personal connection serves only your own requests. A shared connection serves the organisation: everyone, or only the teams an admin chooses. Admins can always add shared connections; other members need the permission.
Credentials are encrypted at rest. Tool results are cached briefly so agent loops do not repeat identical calls, and the facts the model extracts are saved to your organisation's memory with provenance (system, record and link) so they can be checked and superseded.
Host and API key. On Qdrant Cloud copy a key from the cluster's API Keys tab. Self-hosted, use the read_only_api_key or the api_key from the service config. A read-only key is safer and enough for everything but the writes.
Seven-day free trial, then one monthly price per organisation. Connectors, jobs and memory are never metered.
Sign in, add a provider key, create an API key and point your tools at Memtro. It takes about five minutes.
7-day free trial · card on file, nothing charged until it ends · cancel any time