One shared live-data source for your multi-agent system.
When several agents act on the same world, they need to read the same facts. Dynamic Feed is one keyless source every agent in your system can query, returning 87 tools of live data with a signature and provenance stamp on each datapoint.
The call
One keyless POST returns the same signed, provenance-stamped snapshot to every agent that calls it. Batch several tools in a single request so a planner and its workers share one consistent read of the world:
curl -s -X POST https://dynamicfeed.ai/v1/batch \
-H 'Content-Type: application/json' \
-d '{"calls":[
{"tool":"reality_check","args":{"claim":"the latest stable Python is 3.12"}},
{"tool":"service_status","args":{"service":"openai"}},
{"tool":"exploited_vulnerabilities","args":{}}
]}'Sample response
{ "mcpServers": { "dynamic-feed": { "url": "https://dynamicfeed.ai/mcp" } } }Why live data
In a multi-agent system, the failure mode is disagreement: a planner reads one number, a worker reads another, and they make conflicting decisions. A single shared source fixes that. Point every agent at the keyless MCP endpoint (https://dynamicfeed.ai/mcp, Streamable HTTP) or the keyless REST batch endpoint, and they all read from the same place. Each datapoint carries a source, an observed-at timestamp, and a freshness state, so an agent can reason about how stale a value is before acting on it. Each datapoint is also Ed25519-signed. Be precise about what that signature means: it is proof-of-existence and tamper-evidence. It proves Dynamic Feed reported value X at time T and that the bytes have not changed since. It is NOT a claim that X is true, accurate, or safe to act on. Your agents still own judgment; the feed gives them a shared, verifiable, citeable record to judge against. Note there is no live crypto, metal, or stock price conversion here; the macro and rates feeds (interest_rates, treasury_yields, global_macro, china_data) are live economic data, not market prices.
Use it for
- Give a planner and its worker agents one consistent read of the world
- Let agents cite a signed, timestamped fact instead of each guessing
- Add live hazard, status, and CVE awareness to an autonomous workflow
- Share one MCP endpoint across LangChain, LlamaIndex, and CrewAI agents
FAQ
Do all my agents need separate API keys?
No. Both the MCP endpoint and the REST batch endpoint are keyless and free for fair use, so every agent in your system can point at the same URL with no per-agent setup.
How does this keep agents in sync?
They read from one source. Each datapoint returns a source, an observed-at timestamp, and a freshness state, and identical calls return identical signed bytes, so a planner and its workers reason over the same snapshot instead of diverging.
What does the Ed25519 signature actually prove?
It is proof-of-existence and tamper-evidence: that Dynamic Feed reported a given value at a given time and the bytes are unchanged. It does not assert the value is true, accurate, or safe. Your agents still apply judgment.