Skip to content

fix(tavily): updated server instructions to address tavily issues #127

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
paychex-inder merged 3 commits into from
Mar 25, 2026
Merged

fix(tavily): updated server instructions to address tavily issues #127

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
5a0569d
fix(tavily): updated server instructions to address tavily issues
gmarcelino-paychex Mar 23, 2026
9737d24
fix(tavily): enhance schema handling for Tavily tools
gmarcelino-paychex Mar 23, 2026
9905dc8
fix(tavily): revert MCP.js changes
gmarcelino-paychex Mar 24, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion librechat.n1.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -71,7 +71,7 @@ mcpServers:
Authorization: "Bearer ${TAVILY_API_KEY}"
startup: true
chatMenu: false
serverInstructions: "Use Tavily search/news tools only by default. Do NOT use Tavily research or deep-research tools, and do NOT route prompts containing general 'research' wording to those tools. Only use Tavily research/deep-research tools when the user explicitly requests 'deep research'."
serverInstructions: "TAVILY SEARCH PROTOCOL: (1) Never use tavily_research tool unless the user explicitly includes '/deep-research' in their prompt. For all other queries, use tavily_search which has higher rate limits. If /deep-research is requested, use 'mini' model first, and only escalate to 'pro' if the task clearly requires broader multi-topic coverage. (2) Before invoking any Tavily function, validate that the 'query' parameter contains a non-empty string and all parameters match the expected API schema. (3) Execute searches ONE AT A TIME - wait for and verify results from the current search before initiating any additional searches. Never launch multiple parallel search requests. (4) If a search fails or returns errors, analyze the error message, modify the query or parameters, and retry with adjustments. Continue iterating up to 25 attempts to achieve a successful search."
requiresOAuth: false
endpoints:
azureOpenAI:
Expand Down
2 changes: 1 addition & 1 deletion librechat.n2a.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -72,7 +72,7 @@ mcpServers:
Authorization: "Bearer ${TAVILY_API_KEY}"
startup: true
chatMenu: false
serverInstructions: "Use Tavily search/news tools only by default. Do NOT use Tavily research or deep-research tools, and do NOT route prompts containing general 'research' wording to those tools. Only use Tavily research/deep-research tools when the user explicitly requests 'deep research'."
serverInstructions: "TAVILY SEARCH PROTOCOL: (1) Never use tavily_research tool unless the user explicitly includes '/deep-research' in their prompt. For all other queries, use tavily_search which has higher rate limits. If /deep-research is requested, use 'mini' model first, and only escalate to 'pro' if the task clearly requires broader multi-topic coverage. (2) Before invoking any Tavily function, validate that the 'query' parameter contains a non-empty string and all parameters match the expected API schema. (3) Execute searches ONE AT A TIME - wait for and verify results from the current search before initiating any additional searches. Never launch multiple parallel search requests. (4) If a search fails or returns errors, analyze the error message, modify the query or parameters, and retry with adjustments. Continue iterating up to 25 attempts to achieve a successful search."
requiresOAuth: false
endpoints:
azureOpenAI:
Expand Down
2 changes: 1 addition & 1 deletion librechat.prod.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -70,7 +70,7 @@ mcpServers:
Authorization: "Bearer ${TAVILY_API_KEY}"
startup: true
chatMenu: false
serverInstructions: "Use Tavily search/news tools only by default. Do NOT use Tavily research or deep-research tools, and do NOT route prompts containing general 'research' wording to those tools. Only use Tavily research/deep-research tools when the user explicitly requests 'deep research'."
serverInstructions: "TAVILY SEARCH PROTOCOL: (1) Never use tavily_research tool unless the user explicitly includes '/deep-research' in their prompt. For all other queries, use tavily_search which has higher rate limits. If /deep-research is requested, use 'mini' model first, and only escalate to 'pro' if the task clearly requires broader multi-topic coverage. (2) Before invoking any Tavily function, validate that the 'query' parameter contains a non-empty string and all parameters match the expected API schema. (3) Execute searches ONE AT A TIME - wait for and verify results from the current search before initiating any additional searches. Never launch multiple parallel search requests. (4) If a search fails or returns errors, analyze the error message, modify the query or parameters, and retry with adjustments. Continue iterating up to 25 attempts to achieve a successful search."
requiresOAuth: false
endpoints:
azureOpenAI:
Expand Down
Loading