Relative Time Formatting for AI Agents
The tool
Relative Time Formatter provides natural language time representations with support for past and future times, multiple formatting styles (short, long, abbreviated), and detailed time unit breakdowns.
Once your client is connected to the VerveContext MCP server, this appears in its tool list as RelativeTimeFormattingforAIAgents. It is read-only and open-world: it fetches, it never mutates anything on your side, so most clients will call it without asking you to confirm.
{
"name": "RelativeTimeFormattingforAIAgents",
"arguments": {
"timestamp": "1609459200"
}
}What that looks like in a conversation
You don't name the tool — the model picks it. Asking about 1609459200 in the terms this source covers is enough for it to reach for RelativeTimeFormattingforAIAgents on its own.
Connecting
One server URL covers every source in the catalog, including this one. Authorization is OAuth — the client opens a browser once and there is no key to paste into a config file.
{
"mcpServers": {
"vervecontext": {
"url": "https://api.vervecontext.com/v1/mcp"
}
}
}https://api.vervecontext.com/v1/mcpPer-client setup — Claude Desktop, Cursor, VS Code, ChatGPT — is on the MCP setup page.
Arguments
These are the properties on the tool's inputSchema, so a well-behaved client validates them before the call is made.
| Argument | Type | Example | Description |
|---|---|---|---|
timestampRequired | string | 1609459200 | Unix timestamp in seconds or milliseconds |
reference | string | 2024-06-01T00:00:00Z | Reference time to compare against (default: current time) |
style | string | short | Output style One of: short, long, abbreviated. |
What the model gets back
The result carries a structuredContent object matching the tool's declared outputSchema, so a client can read fields without parsing prose. status is "ok" on success and error is null; a null field means the value wasn't available for that input, not that the call failed.
{
"status": "ok",
"error": null,
"data": {
"target_date": "2021-01-01T00:00:00Z",
"reference_date": "2025-12-16T22:28:24.459Z",
"relative_time": "4 years ago",
"is_past": true,
"is_future": false,
"difference_ms": -156464904459,
"primary_unit": "year",
"primary_value": 4,
"all_units": {
"years": 4,
"months": 59,
"weeks": 258,
"days": 1810,
"hours": 43462,
"minutes": 2607748,
"seconds": 156464904,
"milliseconds": 156464904459
},
"style": "short"
}
}Fields
| Field | Type | Example | Description |
|---|---|---|---|
target_date | string | The date being described | |
reference_date | string | The date it is described relative to | |
relative_time | string | Relative time formatted string (e.g., '2 hours ago') | |
is_past | boolean | Whether the target date falls before the reference date | |
is_future | boolean | Whether the target date falls after the reference date | |
difference_ms | number | Gap between the two dates in milliseconds | |
primary_unit | string | Primary time unit used in formatting | |
primary_value | number | Numeric value of the time unit | |
all_units | object | The same gap expressed in every unit | |
└ all_units.years | number | Gap in years | |
└ all_units.months | number | Gap in months | |
└ all_units.weeks | number | Gap in weeks | |
└ all_units.days | number | Gap in days | |
└ all_units.hours | number | Gap in hours | |
└ all_units.minutes | number | Gap in minutes | |
└ all_units.seconds | number | Gap in seconds | |
└ all_units.milliseconds | number | Gap in milliseconds | |
style | string |
Why ground on it
A model can produce something that looks like this answer from its training data, and be confidently out of date or simply wrong. This source returns the current value with a shape you can check, which is the difference between an answer you can cite and one you have to hedge.
Point an evaluation at target_date: it is the field most worth pinning a claim to, and it is either present and current or absent — never plausibly invented.
Failure modes
Errors come back as tool errors with a sentence the model can act on, not a bare status code.
| Status | What it means |
|---|---|
400 / 422 | The arguments didn't validate. The message names the offending one. |
401 | The OAuth session is invalid or expired — reconnect the server. |
403 | Out of credits. Not a bad key — the month's allowance is spent. |
404 | This source isn't part of VerveContext. Check the catalog. |
429 | Brief rate limit. Retrying after a moment succeeds. |
Other ways to use Relative Time Formatting for AI Agents
Set up Relative Time Formatting for AI Agents on VerveContext, or reach the same source a different way. Your VerveContext account and credits work on all of them — one key, one balance.
Frequently asked questions
Do I have to tell the agent to use RelativeTimeFormattingforAIAgents?
No. The tool's name and description are in the model's context once the server is connected, so it selects the tool when the question calls for it. Naming it explicitly works too and is useful when you want to force the call.
Does connecting the server expose every tool at once?
Yes — one connection lists the whole VerveContext catalog. Clients with a tool budget can usually filter the list; the credit cost is per call, so an unused tool costs nothing.
What does a call cost?
1 credit each time the tool actually runs. A model that reasons about the tool without calling it costs nothing.
Is there a REST version of this?
Yes — the same source is available as a plain HTTPS endpoint on APIVerve, linked above. Same data, same credits, same account.