Package GetTDData

Information regarding prices and yields of bonds issued by the Brazilian government can be downloaded from the official open data portal at Tesouro Transparente. However, aggregating all of this historical data into a structured format can be challenging.

Package GetTDData makes importing data from Tesouro Direto simple and fast. All you need to download bond data is the asset identifier (LFT, LTN, NTN-C, NTN-B, NTN-B Principal, NTN-F, NTN-B1, Educa+, RendA+). Function td_get2() downloads directly from the official Tesouro Transparente CKAN open data repository.

Installation

# from CRAN (stable version)
install.package('GetTDData')

# from github (development version)
devtools::install_github('msperlin/GetTDData')

How to use GetTDData

Suppose you need financial data (prices and yields) for a bond of type LTN with a maturity (end of contract) at 2023-01-01. This bullet bond is the most basic debt contract the Brazilian government issues. It does not pay any coupon during its lifetime and pays R$ 1,000 at maturity.

In order to get the data, all you need to do is run the following code in R using td_get2():

library(GetTDData)

assets <- 'LTN'   # Identifier of assets 
first_year <- 2020
last_year <- 2022

df_td <- td_get2(assets,
                 first_year,
                 last_year)
#> 
#> ── Downloading TD data from Tesouro Transparente CKAN
#> 
#> ── Reading TD data
#> ✔ Retrieved 3644 rows of TD data.

Let’s plot the prices to check if the code worked:

library(ggplot2)
library(dplyr)
#> 
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#> 
#>     filter, lag
#> The following objects are masked from 'package:base':
#> 
#>     intersect, setdiff, setequal, union

# filter  LTN 
my_asset_code <- "LTN 010123"

LTN <- df_td %>%
  filter(asset_code  ==  my_asset_code)

p <- ggplot(data = LTN, 
            aes(x = as.Date(ref_date), 
                y = price_bid, 
                color = asset_code)) + 
  geom_line(linewidth = 1) + scale_x_date() + labs(title = '', x = 'Dates')

print(p)

Downloading the Brazilian Yield Curve

The latest version of GetTDData offers function get_yield_curve to download the current Brazilian yield curve directly from Anbima. The yield curve is a tool of financial analysts that show, based on current prices of fixed income instruments, how the market perceives the future real, nominal and inflation returns. You can find more details regarding the use and definition of a yield curve in [Investopedia][https://www.investopedia.com/terms/y/yieldcurve.asp].

df_yield <- get_yield_curve()  
str(df_yield)
#> tibble [103 × 5] (S3: tbl_df/tbl/data.frame)
#>  $ n_biz_days  : num [1:103] 252 252 252 378 378 378 504 504 504 630 ...
#>  $ type        : chr [1:103] "real_return" "nominal_return" "implicit_inflation" "real_return" ...
#>  $ value       : num [1:103] 6.88 13.48 6.17 7.35 13.66 ...
#>  $ ref_date    : Date[1:103], format: "2027-09-13" "2027-09-13" ...
#>  $ current_date: Date[1:103], format: "2026-09-09" "2026-09-09" ...

And we can plot it for the desired result:

library(ggplot2)

p <- ggplot(df_yield, aes(x=ref_date, y = value) ) +
  geom_line(size=1) + geom_point() + facet_grid(~type, scales = 'free') + 
  labs(title = paste0('The current Brazilian Yield Curve '),
       subtitle = paste0('Date: ', df_yield$current_date[1]))     
#> Warning: Using `size` aesthetic for lines was deprecated in ggplot2 3.4.0.
#> ℹ Please use `linewidth` instead.
#> This warning is displayed once per session.
#> Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
#> generated.

print(p)

Using GetTDData as an MCP server

Recent versions of GetTDData can expose a curated set of its functions as Model Context Protocol (MCP) tools, so AI assistants such as OpenCode, Claude Desktop, Claude Code, or Positron Assistant can fetch Brazilian bond data on demand.

Install the optional dependencies:

install.packages(c("mcptools", "ellmer", "jsonlite"))

The command that starts the server is:

Rscript -e "GetTDData::mcp_server()"

Register it with your MCP client. For example, in OpenCode's opencode.json:

{
  "mcp": {
    "servers": {
      "GetTDData": {
        "type": "local",
        "command": ["Rscript", "-e", "GetTDData::mcp_server()"]
      }
    }
  }
}

The server exposes get_td_data, get_td_current, get_yield_curve, list_asset_info, and list_asset_names. You can inspect them from R with names(mcp_tools()).