Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.getbindu.com/llms.txt

Use this file to discover all available pages before exploring further.

This chapter has six steps. Follow them in order.

Step 1 - What you need

You need two things before starting. You may already have them; skim and decide.

Node.js 22+

The gateway is TypeScript; we run it with tsx, no separate build step.

OpenRouter API key

Paid proxy to dozens of models. The gateway uses it for the planner LLM.
node --version    # should print v22.x or higher
Sign up at openrouter.ai, add a few dollars of credit, and copy the key from the API section. It looks like sk-or-v1-<long random string>.
No database required. The gateway is stateless — it holds per-request state in memory for the lifetime of each /plan call and drops it when the call ends. The calling client owns durable history: pass prior turns in the history field and the latest compaction summary in prior_summary on the next call. Old releases required Supabase; that’s gone.

Step 2 - Get the code and install

git clone https://github.com/GetBindu/Bindu
cd Bindu

# Python side - runs the small sample agents we'll call
uv sync --dev --extra agents

# TypeScript side - runs the gateway
cd gateway
npm install
cd ..
The uv sync line uses uv, a fast Python package manager. If you don’t have it:
curl -LsSf https://astral.sh/uv/install.sh | sh

Step 3 - Configure the gateway

Create gateway/.env.local from the template:
cp gateway/.env.example gateway/.env.local
Open it in an editor. Fill in:
gateway/.env.local
# One bearer token the caller must send to talk to the gateway.
# Generate a strong one:
#   openssl rand -base64 32 | tr -d '=' | tr '+/' '-_'
# Paste the output here:
GATEWAY_API_KEY=<paste generated token>

# The planner AI
OPENROUTER_API_KEY=sk-or-v1-<your key>

# Gateway listens here (these are optional — defaults shown)
GATEWAY_PORT=3774
GATEWAY_HOSTNAME=0.0.0.0
And examples/.env (used by the sample Python agents - the file already exists, you just add the key):
examples/.env
OPENROUTER_API_KEY=sk-or-v1-<same key>
What’s a “bearer token”?Think of GATEWAY_API_KEY like the password on a movie ticket booth. Whoever holds this string can ask the gateway to do work on their behalf. The gateway checks it on every request by hashing both sides and comparing the hashes in constant time (so neither a timing nor a length attack can recover the token). Don’t paste it into chat apps or commit it to a public repo. Rotate it when you suspect it leaked.
You may notice old SUPABASE_URL / SUPABASE_SERVICE_ROLE_KEY lines in .env.example. Leave them blank — the gateway no longer reads them. See gateway/src/config/loader.ts:110 for the explicit removal note in the code.

Step 4 - Start one agent

Open a terminal. Start the joke agent — one Python file that answers with jokes. We pin the port to 3773 with BINDU_PORT so it matches the next chapter’s fleet layout (the file’s own default is 5773):
BINDU_PORT=3773 uv run python examples/gateway_test_fleet/joke_agent.py
The boot prints a block of bindufy setup logs. The last line you should see is:
INFO bindu.utils.server_runner: Starting uvicorn server at http://0.0.0.0:3773...
INFO bindu.utils.server_runner: Press Ctrl+C to stop the server gracefully
INFO:     Uvicorn running on http://0.0.0.0:3773 (Press CTRL+C to quit)
Leave that terminal running.

Step 5 - Start the gateway

In a second terminal:
cd gateway
npm run dev
Expected output:
[bindu-gateway] no DID identity configured (set BINDU_GATEWAY_DID_SEED, _AUTHOR, _NAME to enable did_signed peer auth)
[bindu-gateway] listening on http://0.0.0.0:3774
[bindu-gateway] session mode: stateless
The “no DID identity configured” line is expected for now. The DID signing chapter turns on cryptographic signing. session mode: stateless is the only mode in the current gateway — the mode field is kept on the schema for forward-compat but stateful is rejected at boot. Leave this terminal running too.

Step 6 - Ask a question

In a third terminal, load your gateway token into the shell so you don’t have to copy-paste it every time:
set -a && source gateway/.env.local && set +a
Now send the request:
curl -N http://localhost:3774/plan \
  -H "Authorization: Bearer ${GATEWAY_API_KEY}" \
  -H "Content-Type: application/json" \
  -d '{
    "question": "Tell me a joke about databases.",
    "agents": [
      {
        "name": "joke",
        "endpoint": "http://localhost:3773",
        "auth": { "type": "none" },
        "skills": [{ "id": "tell_joke", "description": "Tell a joke" }]
      }
    ]
  }'
The -N flag tells curl not to buffer - you’ll see output appear one line at a time over about 5 seconds.
Expected stream (a few fields like agent_did, agent_did_source, and signatures are elided for readability — they’re documented in the Gateway API reference):
event: session
data: {"session_id":"8f3c4a1e-...","external_session_id":null,"created":true}

event: plan
data: {"plan_id":"b21e7c98-...","session_id":"8f3c4a1e-..."}

event: task.started
data: {"task_id":"call_01H...","agent":"joke","skill":"tell_joke","input":{"input":"Tell me a joke about databases."}}

event: task.artifact
data: {"task_id":"call_01H...","agent":"joke","content":"<remote_content agent=\"joke\" verified=\"unknown\">Why did the database admin break up? Because they had too many relationships!</remote_content>"}

event: task.finished
data: {"task_id":"call_01H...","agent":"joke","state":"completed"}

event: text.delta
data: {"session_id":"8f3c4a1e-...","part_id":"c9d2f60a-...","delta":"Here"}

event: text.delta
data: {"session_id":"8f3c4a1e-...","part_id":"c9d2f60a-...","delta":"'s a joke..."}
... (many more deltas) ...

event: final
data: {"session_id":"8f3c4a1e-...","stop_reason":"stop","usage":{"inputTokens":1130,"outputTokens":52,"totalTokens":1182,"cachedInputTokens":0}}

event: done
data: {}
You made a plan. 🎉

Reading the output line by line

That format is called Server-Sent Events (SSE). It’s plain HTTP, but the server keeps the connection open and writes events one at a time instead of sending one big response at the end. Two parts per event: a label (event: session) and a JSON payload (data: {...}). What each event means, in the order they arrived:
#EventWhat it means
1sessionThe gateway opened an in-memory session for this /plan call. session_id is a per-call handle; external_session_id echoes whatever the client sent (or null). Since the gateway is stateless, this id is not a resumption key — to give the planner context across requests, pass prior turns via the history field on each call.
2planThe planner started its first turn.
3task.startedThe planner decided to call the joke agent. input: {input: "..."} is what it’s sending.
4task.artifactThe agent replied. The text inside <remote_content> is the real answer. That envelope is there so the planner (and you) remember this is untrusted data. The verified attribute takes one of four values — yes, no, unsigned, unknown — and is unknown here because the request didn’t ask the gateway to verify peer signatures. Verification is opt-in per peer via the agent’s trust.verifyDID: true flag (with an optional trust.pinnedDID) — it is independent of auth.type, which controls how the gateway authenticates outbound to the peer. Once trust.verifyDID is on, the label becomes yes (signed and verified), no (signature mismatch), or unsigned (the peer didn’t sign).
5task.finishedThat call is complete.
6text.delta (many)The planner is now writing its own final answer, streamed a word or two at a time. Concatenate them in order (they all share a part_id).
7finalDone. stop_reason: "stop" means “natural end”. usage reports token counts for billing.
8doneLast event. Close the connection.
You may also see one more event on long conversations:
EventWhenMeaning
compaction-summaryMid-stream, zero or one per callThe planner ran out of context window, compacted history into a summary. Persist the summary field client-side and ship it back as prior_summary on the next /plan call.

What’s actually running

You now have three things talking to each other:
The gateway is a coordinator. It doesn’t answer the question itself; it picks an agent, sends the question, gets the reply, writes a final summary using its own planner LLM. It also doesn’t persist anything — when the /plan call ends, the session is gone.
If this is the moment the idea clicks - great. Next we’ll add a second agent so the gateway has a real choice to make: Adding a second agent → Sunflower LogoGateway is stateless - the client owns history.