workflow.md

export CHROMA=http://localhost:8000

1) List tenants (discover defaults)

curl -fsS $CHROMA/api/v2/tenants | jq .

2) Create a tenant (only if you don’t see one yet)

curl -fsS -X POST $CHROMA/api/v2/tenants
-H 'Content-Type: application/json'
-d '{"name":"local"}' | jq .

3) List DBs for that tenant

curl -fsS $CHROMA/api/v2/tenants/local/databases | jq .

4) Create a DB (only if needed)

curl -fsS -X POST $CHROMA/api/v2/tenants/local/databases
-H 'Content-Type: application/json'
-d '{"name":"lore"}' | jq .

5) Create a collection

curl -fsS -X POST $CHROMA/api/v2/tenants/local/databases/lore/collections
-H 'Content-Type: application/json'
-d '{"name":"smoke","metadata":{}}' | jq .

6) Grab the collection id

COL=$( curl -fsS $CHROMA/api/v2/tenants/local/databases/lore/collections
| jq -r '.collections[] | select(.name=="smoke") | .id' ) echo "Collection ID: $COL"

7) Count items (should be 0)

curl -fsS $CHROMA/api/v2/tenants/local/databases/lore/collections/$COL/count | jq .

8) Upsert two vectors (no server embedder? send embeddings explicitly)

curl -fsS -X POST $CHROMA/api/v2/tenants/local/databases/lore/collections/$COL/upsert
-H 'Content-Type: application/json'
-d '{ "ids": ["1","2"], "embeddings": [[0.1,0.2,0.3],[0.0,0.1,0.0]], "metadatas": [{"source":"smoke"},{"source":"smoke"}], "documents": ["hello world","storm goddess pact"] }' | jq .

9) Query nearest neighbor

curl -fsS -X POST $CHROMA/api/v2/tenants/local/databases/lore/collections/$COL/query
-H 'Content-Type: application/json'
-d '{"query_embeddings":[[0.1,0.21,0.31]], "n_results":1, "include":["documents","metadatas","distances"]}' | jq .