Variable Parsing when in LLM Prompts

[Namec999]

Self Checks

• This is only for bug report, if you would like to ask a question, please head to Discussions.
• I have searched for existing issues search for existing issues, including closed ones.
• I confirm that I am using English to submit this report (我已阅读并同意 Language Policy).
• [FOR CHINESE USERS] 请务必使用英文提交 Issue，否则会被关闭。谢谢！:）
• Please do not modify this template :) and fill in all the required fields.

Dify version

0.14.2

Cloud or Self Hosted

Cloud, Self Hosted (Docker), Self Hosted (Source)

Steps to reproduce

when assigning a converastion variable {{#conversation.var1#}} into antother variable,

the use of this variable into any LLM Node Prompte will consider the inner variables as string and not parsing them into variables.

✔️ Expected Behavior

when a variable is used inside a string, stored in another variable or passed within the workflow. it must be parsed as a variabe a not as a string

❌ Actual Behavior

variables in text {{#conversation.testVar#}} or /Context are considered as string, in node LLM prompts

[dosubot]

To ensure that variables used inside a string in LLM prompts are parsed as variables and not as strings when using Dify version 0.14.2, you should use the '{{}}' format for variable interpolation. For example, a variable like '/name' should be written as '{{name}}' to be properly recognized and processed as a variable reference ^[1].

Additionally, Dify supports the use of the Jinja-2 template language for more complex prompts, which allows for lightweight data transformation and logical processing. You can also call out the variable insertion menu by typing / or { to insert special variable blocks or upstream node variables into the prompt as context content ^[2].

To continue talking to Dosu, mention @dosu.

---

^{Help Dosu learn! Give it feedback: Great Response | Irrelevant Answer | Incorrect Sources | Too Verbose | Hallucination | Other}

[Namec999]

i mean i have a variable named var_ONE : "here is the value {{conversation.key}}"

then in an LLM node, i use the variable var_ONE as a prompte.

this will not parse the variable {{conversation.key}} and get it's value, instead, it's just written then as a string