Query Markup Documents - On-device hybrid search for markdown files with BM25, vector search, and LLM reranking
Nothing to run from here
Tinycontext is installed from its publisher's own source and answers where it runs, so this marketplace is not in the path of a single call. There is no address here to send one to, and a panel that pretended otherwise would be showing you an answer we made up. Install it and call it from your own client — the Installation tab has the entry for each one.
6
tools declared
3k
downloads
local
runs on your machine
—
trust score
What it does
Token-light local memory with SQLite hybrid retrieval for MCP agents.
Quickstart
# 1 — add it to Claude Code
claude mcp add --scope user tinycontext-mcp '--' docker run -i --rm docker.io/marcellm01/tinycontext:v0.5.4
# 2 — ask your agent for something these tools do
# save_memories, recall_memories, list_memories and 3 more
Tinycontext costs nothing on mcprush: there is no plan to choose here, no cap to set and nothing mcprush can bill you for.
Publishernot claimed
TIAll 1 server
TinySuiteHQ1 server
1 category
Collected from a public index. Nobody has claimed this account, so nothing here was written by its author — claim it if it is yours.
Installation
Where are you running it?
Every route below installs the same thing and ends at the same approval screen. This one runs on your machine: your client starts Tinycontext as a process under your own user, with your files and your network, so the tool surface below is what it can do to your computer rather than to a server somewhere else. It has not been scanned here, so read the surface, and its source, before you approve it.
Claude Codedeclared by this listing
This is a public server: you run it yourself and this marketplace is not in the path. You need Docker installed and running. Only the required variables are passed into the container. To set an optional one (MCP_TRANSPORT, TINYCONTEXT_CONFIG_PATH, TINYCONTEXT_MEMORY_DB_PATH, TINYCONTEXT_MODELS_DIR, OTEL_SERVICE_NAME, OTEL_EXPORTER_OTLP_ENDPOINT and 2 more), add -e NAME before the image name and give it a value in the entry’s env: an unset -e would clear the default the image sets. Claude Code registers it in one command; --scope user makes it available in every project. The '--' is in quotes so that PowerShell passes it on.
claude mcp add --scope user tinycontext-mcp '--' docker run -i --rm docker.io/marcellm01/tinycontext:v0.5.4Reconnect, or start a new session, and the tools appear in the model’s tool list.
Free on mcprush, and it runs from its publisher’s own image: this marketplace is not in the path, so there is no key of ours to hold and no card to add.
What it asks of you
Nothing is required, but it can take MCP_TRANSPORT (MCP transport to serve: stdio (default), sse, or streamable-http.), TINYCONTEXT_CONFIG_PATH (Path to a context_config.json inside the container.), TINYCONTEXT_MEMORY_DB_PATH (Path to the TinyContext SQLite database.), TINYCONTEXT_MODELS_DIR (Directory used for downloaded ONNX embedding bundles.), OTEL_SERVICE_NAME (OpenTelemetry service name; defaults to tinycontext when telemetry is enabled.), OTEL_EXPORTER_OTLP_ENDPOINT (OTLP collector endpoint for optional OpenTelemetry traces and metrics.), OTEL_EXPORTER_OTLP_PROTOCOL (OTLP protocol: http/protobuf or grpc.), OTEL_EXPORTER_OTLP_HEADERS (Optional OTLP collector headers.) if you want to set them.
What it writes
One config entry your client uses to start the process locally. A local server runs with your file system and your network, which is why it is priced without metering.
What you approve
6 tools, with what each one reads, writes and reaches shown before you agree — the same list on every route above. Read the tool surface.
Tool surface
What the model actually sees. Descriptions are diffed on every release — see version history.
Read from its source code. A server that builds its tool list when it starts can list more tools than this.
Persist one or more concise memories for later recall. Each memory needs content, and defaults to kind='episodic'; use kind='profile' for durable identity/preference facts, which are always attached to every recall_memories response instead of needing a separate lookup.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| memories | array | yes | List of memory objects. Each object must include content, and may set kind to 'profile' (durable identity/preference facts: what to call the user, what they call you, how they like to work) or 'episod |
Retrieve prompt-ready memories within a token budget. With a query, runs hybrid semantic search and labels each memory high, medium, or low relevance. With no query, returns the newest stored memories in chronological order instead (not semantic search, and not date-range aware -- for 'what did we…
| Parameter | Type | Required | Notes |
|---|---|---|---|
| query | string | no | Question or task description to match against memories. Omit for newest-first chronological recall instead of semantic search. |
| top_k | integer | no | Maximum number of memories to return. Defaults to 5 when query is omitted, or the configured recall top_k otherwise. |
Browse stored memories, with pagination and an optional created_at date range. Unlike recall_memories, this does no semantic ranking and no token-budget cutoff -- it's a cheap, deterministic catalog view for questions like 'what did we do last week', for paging through everything stored, or (with…
| Parameter | Type | Required | Notes |
|---|---|---|---|
| kind | string | no | Filter to 'episodic' or 'profile'. Omit for either. |
| since | string | no | Only include memories created at or after this ISO 8601 timestamp/date (e.g. '2026-08-18' or '2026-08-18T00:00:00Z'). Ground this against current_time from a prior recall_memories or list_memories res |
| until | string | no | Only include memories created at or before this ISO 8601 timestamp/date. |
| limit | integer | no | Max entries to return (default 20, capped at 200). |
| offset | integer | no | Number of entries to skip, for pagination. |
| sort | string | no | 'recent' (default) is newest-first. 'stale' instead surfaces the least-recalled, oldest memories first -- good for finding memories worth consolidating with update_memory or removing with delete_memor |
Fetch one stored memory's full content by ref or id, bypassing recall/ranking entirely. Use after list_memories or recall_memories surfaces a ref you want to inspect in full.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| memory_id | string | yes | ref or full id of the memory to fetch, as returned by save_memories, recall_memories, or list_memories. |
Supersede a stored memory with corrected content. Use recall_memories first to find the ref of the memory to correct.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| memory_id | string | yes | ref or full id of the memory to update, as returned by save_memories or recall_memories. |
| content | string | yes | Corrected memory content. |
Permanently delete a single stored memory by ref (or full id). Use recall_memories first to find the ref of the memory to remove.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| memory_id | string | yes | ref or full id of the memory to delete, as returned by save_memories or recall_memories. |
Free
$0 forever
- Every tool, no call limit
- No card, and mcprush charges nothing for it
- Source published under MIT
- Runs on your machine — nothing of it reaches our gateway
- Nothing to cap, because mcprush bills nothing
What counts against your monthly calls
| Tool | Unit | Calls used | Out of the allowance |
|---|
Nothing here is billable: mcprush charges nothing to install Tinycontext or to call it, at any volume.
Call allowance · this installenforced at the gateway
Monthly allowance, in calls
—
—
Alert at 80% · then the tool returns a normal MCP error
Used this month: —
—Grade not published
No rule-by-rule report on file for this release
Two independent axes, because powerful and malicious are different questions. The grade is threat only. The capability level is blast radius, and it is never a penalty on the grade — it is priced as one subtract-only term in the score, where you can see it.
This listing ships as a container image. The scanner reads packages on npm and PyPI and public source; an image is neither, so there is no grade here — not a withheld one, an unmeasured one.
Release history
It installs from a container image rather than from a versioned package, so the command cannot be pinned to a release. The history below is for reading; the current release on file is 0.5.4.
0.5.4
undated
latestNo release note was published with this version.
—
no reviews yet
Who can review
One review per account, from a signed-in account that does not publish this listing. Nothing else is asked — you can say what you think before you install it. Publishers can reply once per review, and a review can be edited or deleted by whoever wrote it.
Sign in to write one
Any signed-in account that does not publish Tinycontext can review it, once — before installing it or after. A review from an account that has installed it is marked as one.
No reviews yet
Nobody has reviewed this listing. The rating on the card is the mean of the reviews written here and nothing else, so there is no rating until somebody writes the first — which takes a signed-in account that does not publish it, and nothing else.