SDK patterns

Copy the wire format.

Memorylayer does not need a heavy client library to be useful. These snippets show the exact calls agents, scripts, and custom dashboards need.

Snippet JSON Playbook JSON API explorer Read docs

ready-to-copy snippets for common clients

operator and agent workflows

required vendor SDKs; HTTP is enough

fetch python curl manifest

01 / Fetch

Plain HTTP first

Use browser or Node fetch with bearer auth and a simple JSON body.

02 / Python

Scriptable workers

Requests examples cover checkpointing, recall, and integration jobs without a vendor SDK.

03 / Curl

Debug the contract

Shell snippets make it easy to test keys, routes, ingest, and bridge calls manually.

04 / Manifest

Generate clients

Public manifests describe routes, tools, arguments, snippets, playbooks, and examples.

HeadersAuthorization: Bearer or X-API-Key

Body{ tool, args } for bridge calls

DiscoveryUse manifests to render clients without hardcoding

Snippets

These are deliberately plain. They match the service contract exactly.

JavaScript fetch

Call the hosted bridge from a Node or browser-side tool runner.

javascript

const response = await fetch(`${MEMORYLAYER_URL}/api/workspaces/${SLUG}/mcp`, {
  method: "POST",
  headers: {
    "Authorization": `Bearer ${MEMORYLAYER_KEY}`,
    "Content-Type": "application/json"
  },
  body: JSON.stringify({
    tool: "recall_context",
    args: { query: "current project state", max_tokens: 1200 }
  })
});

const payload = await response.json();
console.log(payload.result);

Python requests

Use Memorylayer from scripts, workers, or custom agent harnesses.

python

import os
import requests

base = os.environ["MEMORYLAYER_URL"]
slug = os.environ["SLUG"]
key = os.environ["MEMORYLAYER_KEY"]

response = requests.post(
    f"{base}/api/workspaces/{slug}/mcp",
    headers={"Authorization": f"Bearer {key}"},
    json={"tool": "session_checkpoint", "args": {"note": "handoff saved", "limit": 8}},
    timeout=30,
)
response.raise_for_status()
print(response.json()["result"])

Shell recall

Small curl call for debugging keys and tool output.

shell

curl -X POST \
  -H "Authorization: Bearer $MEMORYLAYER_KEY" \
  -H "Content-Type: application/json" \
  -d '{"tool":"recall_hints","args":{"query":"billing work","top_k":5}}' \
  "$MEMORYLAYER_URL/api/workspaces/$SLUG/mcp" 

Agent bootstrap

Fetch workspace-specific URLs, skills, headers, and tool discovery.

shell

curl -H "Authorization: Bearer $MEMORYLAYER_KEY" \
  "$MEMORYLAYER_URL/api/workspaces/$SLUG/bootstrap" 

Batch ingest

Push a small handoff or imported dataset into workspace memory.

shell

curl -X POST \
  -H "Authorization: Bearer $MEMORYLAYER_KEY" \
  -H "Content-Type: application/json" \
  -d '{"source_name":"handoff.md","source_type":"handoff","items":["Shipped manifest endpoints","Next: test onboarding"],"memory_type":"fact"}' \
  "$MEMORYLAYER_URL/api/workspaces/$SLUG/ingest" 

MCP manifest client

Build a tool picker from the public MCP manifest.

javascript

const manifest = await fetch(`${MEMORYLAYER_URL}/api/mcp/manifest`).then((r) => r.json());
for (const group of manifest.tool_groups) {
  console.log(group.name, group.tools.map((tool) => tool.name));
}

Playbooks

Use these as implementation checklists for agent clients and workspace operators.

New agent session

The agent starts with useful state and leaves a handoff behind.

Fetch bootstrap Load resume_context Call get_skills Use recall_context Save session_checkpoint

Repository handoff

A repo can be picked up by another session without reading stale chat logs.

Ingest summary remember_project remember_decision session_handoff export recent

Memory cleanup

Operators can keep the workspace useful instead of letting memories rot.

quality_metrics dedup batch_tag promote/demote status_history

Investigation pivot

A single entity becomes a navigable map of prior work.

search_entities entity_graph recall_related backlinks focus_brief

Client onboarding

A custom client can wire itself without hardcoded docs.

service manifest MCP manifest capability JSON copy SDK snippet test status